Builtin Functions

Numeric Functions
String Functions
Binary Functions
Spatial Functions
Similarity Functions
Tokenizing Functions
Temporal Functions
Object Functions
Aggregate Functions (Array Functions)
Comparison Functions
Type Functions
Conditional Functions
Miscellaneous Functions

The system provides various classes of functions to support operations on numeric, string, spatial, and temporal data. This document explains how to use these functions.

Numeric Functions

abs

Syntax:
```
abs(numeric_value)
```
Computes the absolute value of the argument.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint/float/double value.
Return Value:
- The absolute value of the argument with the same type as the input argument,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error.

Example:

{ "v1": abs(2013), "v2": abs(-4036), "v3": abs(0), "v4": abs(float("-2013.5")), "v5": abs(double("-2013.593823748327284")) };

The expected result is:

{ "v1": 2013, "v2": 4036, "v3": 0, "v4": 2013.5, "v5": 2013.5938237483274 }

acos

Syntax:
```
acos(numeric_value)
```
Computes the arc cosine value of the argument.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint/float/double value.
Return Value:
- the double arc cosine in radians for the argument, if the argument is in the range of -1 (inclusive) to 1 (inclusive),
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error,
- “NaN” for other legitimate numeric values.

Example:

{ "v1": acos(1), "v2": acos(2), "v3": acos(0), "v4": acos(float("0.5")), "v5": acos(double("-0.5")) };

The expected result is:

{ "v1": 0.0, "v2": "NaN", "v3": 1.5707963267948966, "v4": 1.0471975511965979, "v5": 2.0943951023931957 }

asin

Syntax:
```
asin(numeric_value)
```
Computes the arc sine value of the argument.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint/float/double value.
Return Value:
- the double arc sin in radians for the argument, if the argument is in the range of -1 (inclusive) to 1 (inclusive),
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error,
- “NaN” for other legitimate numeric values.

Example:

{ "v1": asin(1), "v2": asin(2), "v3": asin(0), "v4": asin(float("0.5")), "v5": asin(double("-0.5")) };

The expected result is:

{ "v1": 1.5707963267948966, "v2": "NaN", "v3": 0.0, "v4": 0.5235987755982989, "v5": -0.5235987755982989 }

atan

Syntax:
```
atan(numeric_value)
```
Computes the arc tangent value of the argument.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint/float/double value.
Return Value:
- the double arc tangent in radians for the argument,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error.

Example:

{ "v1": atan(1), "v2": atan(2), "v3": atan(0), "v4": atan(float("0.5")), "v5": atan(double("1000")) };

The expected result is:

{ "v1": 0.7853981633974483, "v2": 1.1071487177940904, "v3": 0.0, "v4": 0.4636476090008061, "v5": 1.5697963271282298 }

atan2

Syntax:
```
atan2(numeric_value1, numeric_value2)
```
Computes the arc tangent value of numeric_value2/numeric_value1.
Arguments:
- numeric_value1: a tinyint/smallint/integer/bigint/float/double value,
- numeric_value2: a tinyint/smallint/integer/bigint/float/double value.
Return Value:
- the double arc tangent in radians for numeric_value1 and numeric_value2,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-numeric input value will cause a type error.

Example:

{ "v1": atan2(1, 2), "v2": atan2(0, 4), "v3": atan2(float("0.5"), double("-0.5")) };

The expected result is:

{ "v1": 0.4636476090008061, "v2": 0.0, "v3": 2.356194490192345 }

ceil

Syntax:
```
ceil(numeric_value)
```
Computes the smallest (closest to negative infinity) number with no fractional part that is not less than the value of the argument. If the argument is already equal to mathematical integer, then the result is the same as the argument.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint/float/double value.
Return Value:
- The ceiling value for the given number in the same type as the input argument,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error.

Example:

{
  "v1": ceil(2013),
  "v2": ceil(-4036),
  "v3": ceil(0.3),
  "v4": ceil(float("-2013.2")),
  "v5": ceil(double("-2013.893823748327284"))
};

The expected result is:

{ "v1": 2013, "v2": -4036, "v3": 1.0, "v4": -2013.0, "v5": -2013.0 }

cos

Syntax:
```
cos(numeric_value)
```
Computes the cosine value of the argument.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint/float/double value.
Return Value:
- the double cosine value for the argument,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error.

Example:

{ "v1": cos(1), "v2": cos(2), "v3": cos(0), "v4": cos(float("0.5")), "v5": cos(double("1000")) };

The expected result is:

{ "v1": 0.5403023058681398, "v2": -0.4161468365471424, "v3": 1.0, "v4": 0.8775825618903728, "v5": 0.562379076290703 }

exp

Syntax:
```
exp(numeric_value)
```
Computes e^{numeric_value}.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint/float/double value.
Return Value:
- e^{numeric_value},
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error.

Example:

{ "v1": exp(1), "v2": exp(2), "v3": exp(0), "v4": exp(float("0.5")), "v5": exp(double("1000")) };

The expected result is:

{ "v1": 2.718281828459045, "v2": 7.38905609893065, "v3": 1.0, "v4": 1.6487212707001282, "v5": "Infinity" }

floor

Syntax:
```
floor(numeric_value)
```
Computes the largest (closest to positive infinity) number with no fractional part that is not greater than the value. If the argument is already equal to mathematical integer, then the result is the same as the argument.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint/float/double value.
Return Value:
- The floor value for the given number in the same type as the input argument,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error.

Example:

{
  "v1": floor(2013),
  "v2": floor(-4036),
  "v3": floor(0.8),
  "v4": floor(float("-2013.2")),
  "v5": floor(double("-2013.893823748327284"))
};

The expected result is:

{ "v1": 2013, "v2": -4036, "v3": 0.0, "v4": -2014.0, "v5": -2014.0 }

ln

Syntax:
```
ln(numeric_value)
```
Computes log_enumeric_value.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint/float/double value.
Return Value:
- log_enumeric_value,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error.

Example:

{ "v1": ln(1), "v2": ln(2), "v3": ln(0), "v4": ln(float("0.5")), "v5": ln(double("1000")) };

The expected result is:

{ "v1": 0.0, "v2": 0.6931471805599453, "v3": "-Infinity", "v4": -0.6931471805599453, "v5": 6.907755278982137 }

log

Syntax:
```
log(numeric_value)
```
Computes log₁₀numeric_value.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint/float/double value.
Return Value:
- log₁₀numeric_value,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error.

Example:

{ "v1": log(1), "v2": log(2), "v3": log(0), "v4": log(float("0.5")), "v5": log(double("1000")) };

The expected result is:

{ "v1": 0.0, "v2": 0.3010299956639812, "v3": "-Infinity", "v4": -0.3010299956639812, "v5": 3.0 }

power

Syntax:
```
power(numeric_value1, numeric_value2)
```
Computes numeric_value1^{numeric_value2}.
Arguments:
- numeric_value1: a tinyint/smallint/integer/bigint/float/double value,
- numeric_value2: a tinyint/smallint/integer/bigint/float/double value.
Return Value:
- numeric_value1^{numeric_value2},
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-numeric input value will cause a type error.

Example:

{ "v1": power(1, 2), "v3": power(0, 4), "v4": power(float("0.5"), double("-0.5")) };

The expected result is:

{ "v1": 1, "v3": 0, "v4": 1.4142135623730951 }

round

Syntax:
```
round(numeric_value)
```
Computes the number with no fractional part that is closest (and also closest to positive infinity) to the argument.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint/float/double value.
Return Value:
- The rounded value for the given number in the same type as the input argument,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error.

Example:

{
  "v1": round(2013),
  "v2": round(-4036),
  "v3": round(0.8),
  "v4": round(float("-2013.256")),
  "v5": round(double("-2013.893823748327284"))
};

The expected result is:

{ "v1": 2013, "v2": -4036, "v3": 1.0, "v4": -2013.0, "v5": -2014.0 }

sign

Syntax:
```
sign(numeric_value)
```
Computes the sign of the argument.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint/float/double value.
Return Value:
- the sign (a tinyint) of the argument, -1 for negative values, 0 for 0, and 1 for positive values,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error.

Example:

{ "v1": sign(1), "v2": sign(2), "v3": sign(0), "v4": sign(float("0.5")), "v5": sign(double("-1000")) };

The expected result is:

{ "v1": 1, "v2": 1, "v3": 0, "v4": 1, "v5": -1 }

sin

Syntax:
```
sin(numeric_value)
```
Computes the sine value of the argument.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint/float/double value.
Return Value:
- the double sine value for the argument,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error.

Example:

{ "v1": sin(1), "v2": sin(2), "v3": sin(0), "v4": sin(float("0.5")), "v5": sin(double("1000")) };

The expected result is:

{ "v1": 0.8414709848078965, "v2": 0.9092974268256817, "v3": 0.0, "v4": 0.479425538604203, "v5": 0.8268795405320025 }

sqrt

Syntax:
```
sqrt(numeric_value)
```
Computes the square root of the argument.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint/float/double value.
Return Value:
- the double square root value for the argument,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error.

Example:

{ "v1": sqrt(1), "v2": sqrt(2), "v3": sqrt(0), "v4": sqrt(float("0.5")), "v5": sqrt(double("1000")) };

The expected result is:

{ "v1": 1.0, "v2": 1.4142135623730951, "v3": 0.0, "v4": 0.7071067811865476, "v5": 31.622776601683793 }

tan

Syntax:
```
tan(numeric_value)
```
Computes the tangent value of the argument.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint/float/double value.
Return Value:
- the double tangent value for the argument,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error.

Example:

{ "v1": tan(1), "v2": tan(2), "v3": tan(0), "v4": tan(float("0.5")), "v5": tan(double("1000")) };

The expected result is:

{ "v1": 1.5574077246549023, "v2": -2.185039863261519, "v3": 0.0, "v4": 0.5463024898437905, "v5": 1.4703241557027185 }

trunc

Syntax:
```
trunc(numeric_value, number_digits)
```
Truncates the number to the given number of integer digits to the right of the decimal point (left if digits is negative). Digits is 0 if not given.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint/float/double value,
- number_digits: a tinyint/smallint/integer/bigint value.
Return Value:
- the double tangent value for the argument,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is missing,
- a type error will be raised if:
  - the first argument is any other non-numeric value,
  - the second argument is any other non-tinyint, non-smallint, non-integer, and non-bigint value.

Example:

{ "v1": trunc(1, 1), "v2": trunc(2, -2), "v3": trunc(0.122, 2), "v4": trunc(float("11.52"), -1), "v5": trunc(double("1000.5252"), 3) };

The expected result is:

{ "v1": 1, "v2": 2, "v3": 0.12, "v4": 10.0, "v5": 1000.525 }

round_half_to_even

Syntax:

round_half_to_even(numeric_value, [precision])

Computes the closest numeric value to numeric_value that is a multiple of ten to the power of minus precision. precision is optional and by default value 0 is used.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint/float/double value.
- precision: an optional tinyint/smallint/integer/bigint field representing the number of digits in the fraction of the the result
Return Value:
- The rounded value for the given number in the same type as the input argument,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- a type error will be raised if:
  - the first argument is any other non-numeric value,
  - or, the second argument is any other non-tinyint, non-smallint, non-integer, or non-bigint value.

Example:

{
  "v1": round_half_to_even(2013),
  "v2": round_half_to_even(-4036),
  "v3": round_half_to_even(0.8),
  "v4": round_half_to_even(float("-2013.256")),
  "v5": round_half_to_even(double("-2013.893823748327284")),
  "v6": round_half_to_even(double("-2013.893823748327284"), 2),
  "v7": round_half_to_even(2013, 4),
  "v8": round_half_to_even(float("-2013.256"), 5)
};

The expected result is:

{ "v1": 2013, "v2": -4036, "v3": 1.0, "v4": -2013.0, "v5": -2014.0, "v6": -2013.89, "v7": 2013, "v8": -2013.256 }

Binary Functions

parse_binary

Syntax:

parse_binary(string, encoding)
Creates a binary from an string encoded in encoding format.
Arguments:
- string : an encoded string,
- encoding : a string notation specifies the encoding type of the given string. Currently we support hex and base64 format.
Return Value:
- a binary that is decoded from the given string,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-string input value will cause a type error.
Example:

[ parse_binary(“ABCDEF0123456789”,“hex”), parse_binary(“abcdef0123456789”,“HEX”), parse_binary(‘QXN0ZXJpeAE=’,“base64”) ];
The expected result is:

[ hex(“ABCDEF0123456789”), hex(“ABCDEF0123456789”), hex(“4173746572697801”) ]

print_binary

Syntax:

print_binary(binary, encoding)
Prints a binary to the required encoding string format.
Arguments:
- binary : a binary data need to be printed.
- encoding : a string notation specifies the expected encoding type. Currently we support hex and base64 format.
Return Value:
- a string that represents the encoded format of a binary,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-string input value will cause a type error.

Example:

[ print_binary(hex("ABCDEF0123456789"), "base64"), print_binary(base64("q83vASNFZ4k="), "hex") ]

The expected result are:
```
[ "q83vASNFZ4k=", "ABCDEF0123456789" ]
```

binary_length

Syntax:

binary_length(binary)
Returns the number of bytes storing the binary data.
Arguments:
- binary : a binary value to be checked.
Return Value:
- an bigint that represents the number of bytes,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-binary input value will cause a type error.
Example:
```
binary_length(hex("00AA"))
```
The expected result is:

2

sub_binary

Syntax:

sub_binary(binary, offset[, length])
Returns the sub binary from the given binary based on the given start offset with the optional length.
Arguments:
- binary : a binary to be extracted,
- offset : a tinyint, smallint, integer, or bigint value as the starting offset of the sub binary in binary (starting at 0),
- length : (Optional) a tinyint, smallint, integer, or bigint value as the length of the sub binary.
Return Value:
- a binary that represents the sub binary,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- a type error will be raised if:
  - the first argument is any other non-binary value,
  - or, the second argument is any other non-integer value,
  - or, the third argument is any other non-integer value, if it is present.
Example:
```
sub_binary(hex("AABBCCDD"), 4);
```
The expected result is
```
hex("DD")
```

binary_concat

Syntax:

binary_concat(array)
Concatenates a binary array or multiset into a single binary.
Arguments:
- array : an array or multiset of binaries (could be null or missing) to be concatenated.
Return Value :
- the concatenated binary value,
- missing if the argument is a missing value,
- null if the argument is a null value,
- missing if any element in the input array is missing,
- null if any element in the input array is null but no element in the input array is missing,
- any other non-array input value or non-binary element in the input array will cause a type error.
Example:

binary_concat([hex(“42”), hex(""), hex(‘42’)]);
The expected result is

hex(“4242”)

Spatial Functions

create_point

Syntax:
```
create_point(x, y)
```
Creates the primitive type point using an x and y value.
Arguments:
x : a double that represents the x-coordinate,
y : a double that represents the y-coordinate.
Return Value:
a point representing the ordered pair (x, y),
missing if any argument is a missing value,
null if any argument is a null value but no argument is a missing value,
any other non-double input value will cause a type error.
Example:
```
{ "point": create_point(30.0,70.0) };
```
The expected result is:
```
{ "point": point("30.0,70.0") }
```

create_line

Syntax:
```
create_line(point1, point2)
```
Creates the primitive type line using point1 and point2.
Arguments:
- point1 : a point that represents the start point of the line.
- point2 : a point that represents the end point of the line.
Return Value:
- a spatial line created using the points provided in point1 and point2,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-point input value will cause a type error.

Example:

{ "line": create_line(create_point(30.0,70.0), create_point(50.0,90.0)) };

The expected result is:

{ "line": line("30.0,70.0 50.0,90.0") }

create_rectangle

Syntax:
```
create_rectangle(point1, point2)
```
Creates the primitive type rectangle using point1 and point2.
Arguments:
- point1 : a point that represents the lower_left point of the rectangle.
- point2 : a point that represents the upper_right point of the rectangle.
Return Value:
- a spatial rectangle created using the points provided in point1 and point2,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-point input value will cause a type error.

Example:

{ "rectangle": create_rectangle(create_point(30.0,70.0), create_point(50.0,90.0)) };

The expected result is:

{ "rectangle": rectangle("30.0,70.0 50.0,90.0") }

create_circle

Syntax:
```
create_circle(point, radius)
```
Creates the primitive type circle using point and radius.
Arguments:
- point : a point that represents the center of the circle.
- radius : a double that represents the radius of the circle.
Return Value:
- a spatial circle created using the center point and the radius provided in point and radius.
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- a type error will be raised if:
  - the first argument is any other non-point value,
  - or, the second argument is any other non-double value.

Example:

{ "circle": create_circle(create_point(30.0,70.0), 5.0) }

The expected result is:
```
{ "circle": circle("30.0,70.0 5.0") }
```

create_polygon

Syntax:
```
create_polygon(array)
```
Creates the primitive type polygon using the double values provided in the argument array. Each two consecutive double values represent a point starting from the first double value in the array. Note that at least six double values should be specified, meaning a total of three points.
Arguments:
- array : an array of doubles representing the points of the polygon.
Return Value:
- a polygon, represents a spatial simple polygon created using the points provided in array.
- missing if the argument is a missing value,
- null if the argument is a null value,
- missing if any element in the input array is missing,
- null if any element in the input array is null but no element in the input array is missing,
- any other non-array input value or non-double element in the input array will cause a type error.

Example:

{ "polygon": create_polygon([1.0,1.0,2.0,2.0,3.0,3.0,4.0,4.0]) };

The expected result is:

{ "polygon": polygon("1.0,1.0 2.0,2.0 3.0,3.0 4.0,4.0") }

get_x/get_y

Syntax:
```
get_x(point) or get_y(point)
```
Returns the x or y coordinates of a point point.
Arguments:
- point : a point.
Return Value:
- a double representing the x or y coordinates of the point point,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-point input value will cause a type error.

Example:

{ "x_coordinate": get_x(create_point(2.3,5.0)), "y_coordinate": get_y(create_point(2.3,5.0)) };

The expected result is:

{ "x_coordinate": 2.3, "y_coordinate": 5.0 }

get_points

Syntax:
```
get_points(spatial_object)
```
Returns an ordered array of the points forming the spatial object spatial_object.
Arguments:
- spatial_object : a point, line, rectangle, circle, or polygon.
Return Value:
- an array of the points forming the spatial object spatial_object,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-spatial-object input value will cause a type error.

Example:

get_points(create_polygon([1.0,1.0,2.0,2.0,3.0,3.0,4.0,4.0]))

The expected result is:

[ point("1.0,1.0"), point("2.0,2.0"), point("3.0,3.0"), point("4.0,4.0") ]

get_center/get_radius

Syntax:

get_center(circle_expression) or get_radius(circle_expression)

Returns the center and the radius of a circle circle_expression, respectively.
Arguments:
- circle_expression : a circle.
Return Value:
- a point or double, represent the center or radius of the circle circle_expression.
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-circle input value will cause a type error.

Example:

{
  "circle_radius": get_radius(create_circle(create_point(6.0,3.0), 1.0)),
  "circle_center": get_center(create_circle(create_point(6.0,3.0), 1.0))
};

The expected result is:

{ "circle_radius": 1.0, "circle_center": point("6.0,3.0") }

spatial_distance

Syntax:
```
spatial_distance(point1, point2)
```
Returns the Euclidean distance between point1 and point2.
Arguments:
- point1 : a point.
- point2 : a point.
Return Value:
- a double as the Euclidean distance between point1 and point2.
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-point input value will cause a type error.

Example:

spatial_distance(point("47.44,80.65"), create_point(30.0,70.0));

The expected result is:
```
20.434678857275934
```

spatial_area

Syntax:
```
spatial_area(spatial_2d_expression)
```
Returns the spatial area of spatial_2d_expression.
Arguments:
- spatial_2d_expression : a rectangle, circle, or polygon.
Return Value:
- a double representing the area of spatial_2d_expression.
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-2d-spatial-object will cause a type error.

Example:

spatial_area(create_circle(create_point(0.0,0.0), 5.0));

The expected result is:
```
78.53981625
```

spatial_intersect

Syntax:

spatial_intersect(spatial_object1, spatial_object2)

Checks whether @arg1 and @arg2 spatially intersect each other.
Arguments:
- spatial_object1 : a point, line, rectangle, circle, or polygon.
- spatial_object2 : a point, line, rectangle, circle, or polygon.
Return Value:
- a boolean representing whether spatial_object1 and spatial_object2 spatially overlap with each other,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-spatial-object input value will cause a type error.

Example:

spatial_intersect(point("39.28,70.48"), create_rectangle(create_point(30.0,70.0), create_point(40.0,80.0)));

The expected result is:
```
true
```

spatial_cell

Syntax:

spatial_cell(point1, point2, x_increment, y_increment)

Returns the grid cell that point1 belongs to.
Arguments:
- point1 : a point representing the point of interest that its grid cell will be returned.
- point2 : a point representing the origin of the grid.
- x_increment : a double, represents X increments.
- y_increment : a double, represents Y increments.
Return Value:
- a rectangle representing the grid cell that point1 belongs to,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- a type error will be raised if:
  - the first or second argument is any other non-point value,
  - or, the second or third argument is any other non-double value.

Example:

spatial_cell(point("39.28,70.48"), create_point(20.0,50.0), 5.5, 6.0);

The expected result is:
```
rectangle("36.5,68.0 42.0,74.0");
```

Similarity Functions

AsterixDB supports queries with different similarity functions, including edit distance and Jaccard.

edit_distance

Syntax:

edit_distance(expression1, expression2)

Returns the edit distance of expression1 and expression2.
Arguments:
- expression1 : a string or a homogeneous array of a comparable item type.
- expression2 : The same type as expression1.
Return Value:
- an bigint that represents the edit distance between expression1 and expression2,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-string input value will cause a type error.
Note: an n_gram index can be utilized for this function.

Example:

edit_distance("SuzannaTillson", "Suzanna Tilson");

The expected result is:
```
2
```

edit_distance_check

Syntax:

edit_distance_check(expression1, expression2, threshold)

Checks whether the edit distance of expression1 and expression2 is within a given threshold.
Arguments:
- expression1 : a string or a homogeneous array of a comparable item type.
- expression2 : The same type as expression1.
- threshold : a bigint that represents the distance threshold.
Return Value:
- an array with two items:
  - The first item contains a boolean value representing whether the edit distance of expression1 and expression2 is within the given threshold.
  - The second item contains an integer that represents the edit distance of expression1 and expression2 if the first item is true.
  - If the first item is false, then the second item is set to 2147483647.
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- a type error will be raised if:
  - the first or second argument is any other non-string value,
  - or, the third argument is any other non-bigint value.
Note: an n_gram index can be utilized for this function.
Example:
```
edit_distance_check("happy","hapr",2);
```
The expected result is:
```
[ true, 2 ]
```

edit_distance_contains

Syntax:

edit_distance_contains(expression1, expression2, threshold)

Checks whether expression1 contains expression2 with an edit distance within a given threshold.
Arguments:
- expression1 : a string or a homogeneous array of a comparable item type.
- expression2 : The same type as expression1.
- threshold : a bigint that represents the distance threshold.
Return Value:
- an array with two items:
  - The first item contains a boolean value representing whether expression1 can contain expression2.
  - The second item contains an integer that represents the required edit distance for expression1 to contain expression2 if the first item is true.
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- a type error will be raised if:
  - the first or second argument is any other non-string value,
  - or, the third argument is any other non-bigint value.
Note: an n_gram index can be utilized for this function.

Example:

edit_distance_contains("happy","hapr",2);

The expected result is:
```
[ true, 1 ]
```

similarity_jaccard

Syntax:
```
similarity_jaccard(array1, array2)
```
Returns the Jaccard similarity of array1 and array2.
Arguments:
- array1 : an array or multiset.
- array2 : an array or multiset.
Return Value:
- a float that represents the Jaccard similarity of array1 and array2,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- missing if any element in any input array is missing,
- null if any element in any input array is null but no element in the input array is missing,
- any other non-array input value or non-integer element in any input array will cause a type error.
Note: a keyword index can be utilized for this function.

Example:

similarity_jaccard([1,5,8,9], [1,5,9,10]);

The expected result is:
```
0.6
```

similarity_jaccard_check

Syntax:

similarity_jaccard_check(array1, array2, threshold)

Checks whether array1 and array2 have a Jaccard similarity greater than or equal to threshold. Again, the “check” version of Jaccard is faster than the “non_check” version.
Arguments:
- array1 : an array or multiset.
- array2 : an array or multiset.
- threshold : a double that represents the similarity threshold.
Return Value:
- an array with two items:
  - The first item contains a boolean value representing whether array1 and array2 are similar.
  - The second item contains a float that represents the Jaccard similarity of array1 and array2 if it is greater than or equal to the threshold, or 0 otherwise.
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- missing if any element in any input array is missing,
- null if any element in any input array is null but no element in the input array is missing,
- a type error will be raised if:
  - the first or second argument is any other non-array value,
    - or, the third argument is any other non-double value.
Note: a keyword index can be utilized for this function.

Example:

similarity_jaccard_check([1,5,8,9], [1,5,9,10], 0.6);

The expected result is:
```
[ false, 0.0 ]
```

Tokenizing Functions

word_tokens

Syntax:
```
word_tokens(string)
```
Returns an array of word tokens of string using non_alphanumeric characters as delimiters.
Arguments:
- string : a string that will be tokenized.
Return Value:
- an array of string word tokens,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-string input value will cause a type error.

Example:

word_tokens("I like the phone, awesome!");

The expected result is:

[ "i", "like", "the", "phone", "awesome" ]

Temporal Functions

get_year/get_month/get_day/get_hour/get_minute/get_second/get_millisecond

Syntax:

get_year/get_month/get_day/get_hour/get_minute/get_second/get_millisecond(temporal_value)

Accessors for accessing fields in a temporal value
Arguments:
- temporal_value : a temporal value represented as one of the following types: date, datetime, time, and duration.
Return Value:
- an bigint value representing the field to be extracted,
- missing if the argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-interval input value will cause a type error.

Example:

{
  "year": get_year(date("2010-10-30")),
  "month": get_month(datetime("1987-11-19T23:49:23.938")),
  "day": get_day(date("2010-10-30")),
  "hour": get_hour(time("12:23:34.930+07:00")),
  "min": get_minute(duration("P3Y73M632DT49H743M3948.94S")),
  "second": get_second(datetime("1987-11-19T23:49:23.938")),
  "ms": get_millisecond(duration("P3Y73M632DT49H743M3948.94S"))
};

The expected result is:

{ "year": 2010, "month": 11, "day": 30, "hour": 5, "min": 28, "second": 23, "ms": 94 }

adjust_datetime_for_timezone

Syntax:

adjust_datetime_for_timezone(datetime, string)

Adjusts the given datetime datetime by applying the timezone information string.
Arguments:
- datetime : a datetime value to be adjusted.
- string : a string representing the timezone information.
Return Value:
- a string value representing the new datetime after being adjusted by the timezone information,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- a type error will be raised if:
  - the first argument is any other non-datetime value,
  - or, the second argument is any other non-string value.

Example:

adjust_datetime_for_timezone(datetime("2008-04-26T10:10:00"), "+08:00");

The expected result is:
```
"2008-04-26T18:10:00.000+08:00"
```

adjust_time_for_timezone

Syntax:
```
adjust_time_for_timezone(time, string)
```
Adjusts the given time time by applying the timezone information string.
Arguments:
- time : a time value to be adjusted.
- string : a string representing the timezone information.
Return Value:
- a string value representing the new time after being adjusted by the timezone information,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- a type error will be raised if:
  - the first argument is any other non-time value,
  - or, the second argument is any other non-string value.

Example:

adjust_time_for_timezone(get_time_from_datetime(datetime("2008-04-26T10:10:00")), "+08:00");

The expected result is:
```
"18:10:00.000+08:00"
```

calendar_duration_from_datetime

Syntax:

calendar_duration_from_datetime(datetime, duration_value)

Gets a user_friendly representation of the duration duration_value based on the given datetime datetime.
Arguments:
- datetime : a datetime value to be used as the reference time point.
- duration_value : a duration value to be converted.
Return Value:
- a duration value with the duration as duration_value but with a user_friendly representation,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- a type error will be raised if:
  - the first argument is any other non-datetime value,
  - or, the second argument is any other non-duration input value.

Example:

calendar_duration_from_datetime(
      datetime("2016-03-26T10:10:00"),
      datetime("2016-03-26T10:10:00") - datetime("2011-01-01T00:00:00")
);

The expected result is:
```
duration("P5Y2M24DT10H10M")
```

get_year_month_duration/get_day_time_duration

Syntax:

get_year_month_duration/get_day_time_duration(duration_value)

Extracts the correct duration subtype from duration_value.
Arguments:
- duration_value : a duration value to be converted.
Return Value:
- a year_month_duration value or a day_time_duration value,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-duration input value will cause a type error.

Example:

get_year_month_duration(duration("P12M50DT10H"));

The expected result is:
```
year_month_duration("P1Y")
```

months_from_year_month_duration/milliseconds_from_day_time_duration

Syntax:

months_from_year_month_duration/milliseconds_from_day_time_duration(duration_value)

Extracts the number of months or the number of milliseconds from the duration subtype.
Arguments:
- duration_value : a duration of the correct subtype.
Return Value:
- an bigint representing the number or months/milliseconds,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-duration input value will cause a type error.

Example:

months_from_year_month_duration(get_year_month_duration(duration("P5Y7MT50M")));

The expected result is:
```
67
```

duration_from_months/duration_from_ms

Syntax:

duration_from_months/duration_from_ms(number_value)

Creates a duration from number_value.
Arguments:
- number_value : a bigint representing the number of months/milliseconds
Return Value:
- a duration containing number_value value for months/milliseconds,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-duration input value will cause a type error.
Example:
```
duration_from_months(8);
```
The expected result is:
```
duration("P8M")
```

duration_from_interval

Syntax:
```
duration_from_interval(interval_value)
```
Creates a duration from interval_value.
Arguments:
- interval_value : an interval value
Return Value:
- a duration representing the time in the interval_value
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-duration input value will cause a type error.

Example:

{
  "dr1" : duration_from_interval(interval(date("2010-10-30"), date("2010-12-21"))),
  "dr2" : duration_from_interval(interval(datetime("2012-06-26T01:01:01.111"), datetime("2012-07-27T02:02:02.222"))),
  "dr3" : duration_from_interval(interval(time("12:32:38"), time("20:29:20"))),
  "dr4" : duration_from_interval(null)
};

The expected result is:

{
  "dr1": day_time_duration("P52D"),
  "dr2": day_time_duration("P31DT1H1M1.111S"),
  "dr3": day_time_duration("PT7H56M42S"),
  "dr4": null
}

current_date

Syntax:
```
current_date()
```
Gets the current date.
Arguments: None
Return Value:
- a date value of the date when the function is called.

current_time

Syntax:
```
current_time()
```
Get the current time
Arguments: None
Return Value:
- a time value of the time when the function is called.

current_datetime

Syntax:
```
current_datetime()
```
Get the current datetime
Arguments: None
Return Value:
- a datetime value of the datetime when the function is called.

get_date_from_datetime

Syntax:
```
get_date_from_datetime(datetime)
```
Gets the date value from the given datetime value datetime.
Arguments:
- datetime: a datetime value to be extracted from.
Return Value:
- a date value from the datetime,
- any other non-datetime input value will cause a type error.

get_time_from_datetime

Syntax:
```
get_time_from_datetime(datetime)
```
Get the time value from the given datetime value datetime
Arguments:
- datetime: a datetime value to be extracted from.
Return Value:
- a time value from the datetime.
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-datetime input value will cause a type error.

Example:

get_time_from_datetime(datetime("2016-03-26T10:10:00"));

The expected result is:
```
time("10:10:00.000Z")
```

day_of_week

Syntax:
```
day_of_week(date)
```
Finds the day of the week for a given date (1_7)
Arguments:
- date: a date value (Can also be a datetime)
Return Value:
- an tinyint representing the day of the week (1_7),
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-date input value will cause a type error.

Example:

day_of_week(datetime("2012-12-30T12:12:12.039Z"));

The expected result is:
```
7
```

date_from_unix_time_in_days

Syntax:

date_from_unix_time_in_days(numeric_value)

Gets a date representing the time after numeric_value days since 1970_01_01.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint value representing the number of days.
Return Value:
- a date value as the time after numeric_value days since 1970-01-01,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error.

datetime_from_unix_time_in_ms

Syntax:

datetime_from_unix_time_in_ms(numeric_value)

Gets a datetime representing the time after numeric_value milliseconds since 1970_01_01T00:00:00Z.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint value representing the number of milliseconds.
Return Value:
- a datetime value as the time after numeric_value milliseconds since 1970-01-01T00:00:00Z,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error.

datetime_from_unix_time_in_secs

Syntax:

datetime_from_unix_time_in_secs(numeric_value)

Gets a datetime representing the time after numeric_value seconds since 1970_01_01T00:00:00Z.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint value representing the number of seconds.
Return Value:
- a datetime value as the time after numeric_value seconds since 1970_01_01T00:00:00Z,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error.

datetime_from_date_time

Syntax:

datetime_from_date_time(date,time)

Gets a datetime representing the combination of date and time
- Arguments:
- date: a date value
- time a time value
Return Value:
- a datetime value by combining date and time,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- a type error will be raised if
  - the first argument is any other non-date value,
  - or, the second argument is any other non-time value.

time_from_unix_time_in_ms

Syntax:

time_from_unix_time_in_ms(numeric_value)

Gets a time representing the time after numeric_value milliseconds since 00:00:00.000Z.
Arguments:
- numeric_value: a tinyint/smallint/integer/bigint value representing the number of milliseconds.
Return Value:
- a time value as the time after numeric_value milliseconds since 00:00:00.000Z,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-numeric input value will cause a type error.

Example:

{
  "date": date_from_unix_time_in_days(15800),
  "datetime": datetime_from_unix_time_in_ms(1365139700000),
  "time": time_from_unix_time_in_ms(3748)
};

The expected result is:

{ "date": date("2013-04-05"), "datetime": datetime("2013-04-05T05:28:20.000Z"), "time": time("00:00:03.748Z") }

unix_time_from_date_in_days

Syntax:

unix_time_from_date_in_days(date_value)

Gets an integer value representing the number of days since 1970_01_01 for date_value.
Arguments:
- date_value: a date value.
Return Value:
- a bigint value representing the number of days,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-date input value will cause a type error.

unix_time_from_datetime_in_ms

Syntax:

unix_time_from_datetime_in_ms(datetime_value)

Gets an integer value representing the time in milliseconds since 1970_01_01T00:00:00Z for datetime_value.
Arguments:
- datetime_value : a datetime value.
Return Value:
- a bigint value representing the number of milliseconds,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-datetime input value will cause a type error.

unix_time_from_datetime_in_secs

Syntax:

unix_time_from_datetime_in_secs(datetime_value)

Gets an integer value representing the time in seconds since 1970_01_01T00:00:00Z for datetime_value.
Arguments:
- datetime_value : a datetime value.
Return Value:
- a bigint value representing the number of seconds,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-datetime input value will cause a type error.

unix_time_from_time_in_ms

Syntax:
```
unix_time_from_time_in_ms(time_value)
```
Gets an integer value representing the time the milliseconds since 00:00:00.000Z for time_value.
Arguments:
- time_value : a time value.
Return Value:
- a bigint value representing the number of milliseconds,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-datetime input value will cause a type error.

Example:

{
  "date": date_from_unix_time_in_days(15800),
  "datetime": datetime_from_unix_time_in_ms(1365139700000),
  "time": time_from_unix_time_in_ms(3748)
}

The expected result is:

{ "date": date("2013-04-05"), "datetime": datetime("2013-04-05T05:28:20.000Z"), "time": time("00:00:03.748Z") }

parse_date/parse_time/parse_datetime

Syntax:

parse_date/parse_time/parse_datetime(date,formatting_expression)

Creates a date/time/date_time value by treating date with formatting formatting_expression
Arguments:
- date: a string value representing the date/time/datetime.
- formatting_expression a string value providing the formatting for date_expression.Characters used to create date expression:
- h hours
- m minutes
- s seconds
- n milliseconds
- a am/pm
- z timezone
- Y year
- M month
- D day
- W weekday
- _, ', /, ., ,, T seperators for both time and date
Return Value:
- a date/time/date_time value corresponding to date,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- a type error will be raised if:
- the first argument is any other non-date value,
- the second argument is any other non-string value.
Example:
```
parse_time("30:30","m:s");
```
The expected result is:
```
time("00:30:30.000Z")
```

print_date/print_time/print_datetime

Syntax:

print_date/print_time/print_datetime(date,formatting_expression)

Creates a string representing a date/time/date_time value of the date using the formatting formatting_expression
Arguments:
- date: a date/time/datetime value.
- formatting_expression a string value providing the formatting for date_expression. Characters used to create date expression:
- h hours
- m minutes
- s seconds
- n milliseconds
- a am/pm
- z timezone
- Y year
- M month
- D day
- W weekday
- _, ', /, ., ,, T seperators for both time and date
Return Value:
- a string value corresponding to date,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- a type error will be raised if:
  - the first argument is any other non-date value,
  - the second argument is any other non-string value.

Example:

print_time(time("00:30:30.000Z"),"m:s");

The expected result is:
```
"30:30"
```

get_interval_start, get_interval_end

Syntax:

get_interval_start/get_interval_end(interval)

Gets the start/end of the given interval.
Arguments:
- interval: the interval to be accessed.
Return Value:
- a time, date, or datetime (depending on the time instances of the interval) representing the starting or ending time,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-interval value will cause a type error.

Example:

{
  "start": get_interval_start(interval_start_from_date("1984-01-01", "P1Y")),
  "end": get_interval_end(interval_start_from_date("1984-01-01", "P1Y"))
};

The expected result is:

{ "start": date("1984_01_01"), "end": date("1985_01_01") }

get_interval_start_date/get_interval_start_datetimeget_interval_start_time, get_interval_end_date/get_interval_end_datetime/get_interval_end_time

Syntax:

get_interval_start_date/get_interval_start_datetime/get_interval_start_time/get_interval_end_date/get_interval_end_datetime/get_interval_end_time(interval)

Gets the start/end of the given interval for the specific date/datetime/time type.
Arguments:
- interval: the interval to be accessed.
Return Value:
- a time, date, or datetime (depending on the function) representing the starting or ending time,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-interval value will cause a type error.

Example:

{
  "start1": get_interval_start_date(interval_start_from_date("1984-01-01", "P1Y")),
  "end1": get_interval_end_date(interval_start_from_date("1984-01-01", "P1Y")),
  "start2": get_interval_start_datetime(interval_start_from_datetime("1984-01-01T08:30:00.000", "P1Y1H")),
  "end2": get_interval_end_datetime(interval_start_from_datetime("1984-01-01T08:30:00.000", "P1Y1H")),
  "start3": get_interval_start_time(interval_start_from_time("08:30:00.000", "P1H")),
  "end3": get_interval_end_time(interval_start_from_time("08:30:00.000", "P1H"))
};

The expected result is:

{
  "start1": date("1984-01-01"),
  "end1": date("1985-01-01"),
  "start2": datetime("1984-01-01T08:30:00.000Z"),
  "end2": datetime("1985-01-01T09:30:00.000Z"),
  "start3": time("08:30:00.000Z"),
  "end3": time("09:30:00.000Z")
}

get_overlapping_interval

Syntax:

get_overlapping_interval(interval1, interval2)

Gets the start/end of the given interval for the specific date/datetime/time type.
Arguments:
- interval1: an interval value
- interval2: an interval value
Return Value:
- an interval that is overlapping interval1 and interval2. If interval1 and interval2 do not overlap null is returned. Note each interval must be of the same type.
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-interval input value will cause a type error.

Example:

{ "overlap1": get_overlapping_interval(interval(time("11:23:39"), time("18:27:19")), interval(time("12:23:39"), time("23:18:00"))),
  "overlap2": get_overlapping_interval(interval(time("12:23:39"), time("18:27:19")), interval(time("07:19:39"), time("09:18:00"))),
  "overlap3": get_overlapping_interval(interval(date("1980-11-30"), date("1999-09-09")), interval(date("2013-01-01"), date("2014-01-01"))),
  "overlap4": get_overlapping_interval(interval(date("1980-11-30"), date("2099-09-09")), interval(date("2013-01-01"), date("2014-01-01"))),
  "overlap5": get_overlapping_interval(interval(datetime("1844-03-03T11:19:39"), datetime("2000-10-30T18:27:19")), interval(datetime("1989-03-04T12:23:39"), datetime("2009-10-10T23:18:00"))),
  "overlap6": get_overlapping_interval(interval(datetime("1989-03-04T12:23:39"), datetime("2000-10-30T18:27:19")), interval(datetime("1844-03-03T11:19:39"), datetime("1888-10-10T23:18:00")))
};

The expected result is:

{ "overlap1": interval(time("12:23:39.000Z"), time("18:27:19.000Z")),
  "overlap2": null,
  "overlap3": null,
  "overlap4": interval(date("2013-01-01"), date("2014_01_01")),
  "overlap5": interval(datetime("1989-03-04T12:23:39.000Z"), datetime("2000-10-30T18:27:19.000Z")),
  "overlap6": null
}

interval_bin

Syntax:

interval_bin(time_to_bin, time_bin_anchor, duration_bin_size)

Returns the interval value representing the bin containing the time_to_bin value.
Arguments:
- time_to_bin: a date/time/datetime value representing the time to be binned.
- time_bin_anchor: a date/time/datetime value representing an anchor of a bin starts. The type of this argument should be the same as the first time_to_bin argument.
- duration_bin_size: the duration value representing the size of the bin, in the type of year_month_duration or day_time_duration. The type of this duration should be compatible with the type of time_to_bin, so that the arithmetic operation between time_to_bin and duration_bin_size is well_defined. Currently AsterixDB supports the following arithmetic operations:
  - datetime +|_ year_month_duration
  - datetime +|_ day_time_duration
  - date +|_ year_month_duration
  - date +|_ day_time_duration
  - time +|_ day_time_duration
Return Value:
- a interval value representing the bin containing the time_to_bin value. Note that the internal type of this interval value should be the same as the time_to_bin type,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- a type error will be raised if:
  - the first argument or the second argument is any other non-date/non-time/non-datetime value,
  - or, the second argument is any other non-year_month_duration/non-day_time_duration value.

Example:

{
  "bin1": interval_bin(date("2010-10-30"), date("1990-01-01"), year_month_duration("P1Y")),
  "bin2": interval_bin(datetime("1987-11-19T23:49:23.938"), datetime("1990-01-01T00:00:00.000Z"), year_month_duration("P6M")),
  "bin3": interval_bin(time("12:23:34.930+07:00"), time("00:00:00"), day_time_duration("PT1M")),
  "bin4": interval_bin(datetime("1987-11-19T23:49:23.938"), datetime("2013-01-01T00:00:00.000"), day_time_duration("PT24H"))
};

The expected result is:

{
  "bin1": interval(date("2010-01-01"),date("2011-01-01")),
  "bin2": interval(datetime("1987-07-01T00:00:00.000Z"), datetime("1988-01-01T00:00:00.000Z")),
  "bin3": interval(time("05:23:00.000Z"), time("05:24:00.000Z")),
  "bin4": interval(datetime("1987-11-19T00:00:00.000Z"), datetime("1987-11-20T00:00:00.000Z"))
}

interval_start_from_date/time/datetime

Syntax:

interval_start_from_date/time/datetime(date/time/datetime, duration)

Construct an interval value by the given starting date/time/datetime and the duration that the interval lasts.
Arguments:
- date/time/datetime: a string representing a date, time or datetime, or a date/time/datetime value, representing the starting time point.
- duration: a string or duration value representing the duration of the interval. Note that duration cannot be negative value.
Return Value:
- an interval value representing the interval starting from the given time point with the length of duration,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- a type error will be raised if:
  - the first argument or the second argument is any other non-date/non-time/non-datetime value,
  - or, the second argument is any other non-duration value.

Example:

{
  "interval1": interval_start_from_date("1984-01-01", "P1Y"),
  "interval2": interval_start_from_time(time("02:23:28.394"), "PT3H24M"),
  "interval3": interval_start_from_datetime("1999-09-09T09:09:09.999", duration("P2M30D"))
};

The expectecd result is:

{
  "interval1": interval(date("1984-01-01"), date("1985-01-01")),
  "interval2": interval(time("02:23:28.394Z"), time("05:47:28.394Z")),
  "interval3": interval(datetime("1999-09-09T09:09:09.999Z"), datetime("1999-12-09T09:09:09.999Z"))
}

overlap_bins

Return Value:
- a interval value representing the bin containing the time_to_bin value. Note that the internal type of this interval value should be the same as the time_to_bin type.

Syntax:

overlap_bins(interval, time_bin_anchor, duration_bin_size)

Returns an ordered list of interval values representing each bin that is overlapping the interval.
Arguments:
- interval: an interval value
- time_bin_anchor: a date/time/datetime value representing an anchor of a bin starts. The type of this argument should be the same as the first time_to_bin argument.
- duration_bin_size: the duration value representing the size of the bin, in the type of year_month_duration or day_time_duration. The type of this duration should be compatible with the type of time_to_bin, so that the arithmetic operation between time_to_bin and duration_bin_size is well_defined. Currently AsterixDB supports the following arithmetic operations:
  - datetime +|_ year_month_duration
  - datetime +|_ day_time_duration
  - date +|_ year_month_duration
  - date +|_ day_time_duration
  - time +|_ day_time_duration
Return Value:
- a ordered list of interval values representing each bin that is overlapping the interval. Note that the internal type as time_to_bin and duration_bin_size.
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- a type error will be raised if:
  - the first arugment is any other non-interval value,
  - or, the second argument is any other non-date/non-time/non-datetime value,
  - or, the second argument is any other non-year_month_duration/non-day_time_duration value.

Example:

{
  "timebins": overlap_bins(interval(time("17:23:37"), time("18:30:21")), time("00:00:00"), day_time_duration("PT30M")),
  "datebins": overlap_bins(interval(date("1984-03-17"), date("2013-08-22")), date("1990-01-01"), year_month_duration("P10Y")),
  "datetimebins": overlap_bins(interval(datetime("1800-01-01T23:59:48.938"), datetime("2015-07-26T13:28:30.218")),
                              datetime("1900-01-01T00:00:00.000"), year_month_duration("P100Y"))
};

The expected result is:

{
  "timebins": [
                interval(time("17:00:00.000Z"), time("17:30:00.000Z")),
                interval(time("17:30:00.000Z"), time("18:00:00.000Z")),
                interval(time("18:00:00.000Z"), time("18:30:00.000Z")),
                interval(time("18:30:00.000Z"), time("19:00:00.000Z"))
              ],
  "datebins": [
                interval(date("1980-01-01"), date("1990-01-01")),
                interval(date("1990-01-01"), date("2000-01-01")),
                interval(date("2000-01-01"), date("2010-01-01")),
                interval(date("2010-01-01"), date("2020-01-01"))
              ],
  "datetimebins": [
                    interval(datetime("1800-01-01T00:00:00.000Z"), datetime("1900-01-01T00:00:00.000Z")),
                    interval(datetime("1900-01-01T00:00:00.000Z"), datetime("2000-01-01T00:00:00.000Z")),
                    interval(datetime("2000-01-01T00:00:00.000Z"), datetime("2100-01-01T00:00:00.000Z"))
                   ]
};

interval_before, interval_after

Syntax:

interval_before(interval1, interval2)
interval_after(interval1, interval2)

These two functions check whether an interval happens before/after another interval.
Arguments:
- interval1, interval2: two intervals to be compared
Return Value:
- a boolean value. Specifically, interval_before(interval1, interval2) is true if and only if interval1.end < interval2.start, and interval_after(interval1, interval2) is true if and only if interval1.start > interval2.end.
- missing if the argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-interval input value will cause a type error.

Examples:

{
  "interval_before": interval_before(interval(date("2000-01-01"), date("2005-01-01")),
                                     interval(date("2005-05-01"), date("2012-09-09"))),
  "interval_after": interval_after(interval(date("2005-05-01"), date("2012-09-09")),
                                   interval(date("2000-01-01"), date("2005-01-01")))
};

The expected result is:

{ "interval_before": true, "interval_after": true }

interval_covers, interval_covered_by

Syntax:

interval_covers(interval1, interval2)
interval_covered_by(interval1, interval2)

These two functions check whether one interval covers the other interval.
Arguments:
- interval1, interval2: two intervals to be compared
Return Value:
- a boolean value. Specifically, interval_covers(interval1, interval2) is true if and only if
  
  interval1.start <= interval2.start AND interval1.end >= interval2.end
  
  interval_covered_by(interval1, interval2) is true if and only if
  
  interval2.start <= interval1.start AND interval2.end >= interval1.end
- missing if the argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-interval input value will cause a type error.

Examples:

{
  "interval_covers": interval_covers(interval(date("2000-01-01"), date("2005-01-01")),
                                     interval(date("2000-03-01"), date("2004-09-09"))),
  "interval_covered_by": interval_covered_by(interval(date("2006-08-01"), date("2007-03-01")),
                                             interval(date("2004-09-10"), date("2012-08-01")))
};

The expected result is:

{ "interval_covers": true, "interval_covered_by": true }

interval_overlaps, interval_overlapped_by

Syntax:

interval_overlaps(interval1, interval2)
interval_overlapped_by(interval1, interval2)

These functions check whether two intervals overlap with each other.
Arguments:
- interval1, interval2: two intervals to be compared
Return Value:
- a boolean value. Specifically, interval_overlaps(interval1, interval2) is true if and only if
  interval1.start < interval2.start AND interval2.end > interval1.end AND interval1.end > interval2.start
interval_overlapped_by(interval1, interval2) is true if and only if
```
interval2.start < interval1.start
AND interval1.end > interval2.end
AND interval2.end > interval1.start
```
- missing if the argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-interval input value will cause a type error.
Note that interval_overlaps and interval_overlapped_by are following the Allen’s relations on the definition of overlap.

Examples:

{
  "overlaps": interval_overlaps(interval(date("2000-01-01"), date("2005-01-01")),
                                interval(date("2004-05-01"), date("2012-09-09"))),
  "overlapped_by": interval_overlapped_by(interval(date("2006-08-01"), date("2007-03-01")),
                                          interval(date("2004-05-01"), date("2012-09-09"))))
};

The expected result is:

{ "overlaps": true, "overlapped_by": true }

interval_overlapping

Note that interval_overlapping is not an Allen’s Relation, but syntactic sugar we added for the case that the intersect of two intervals is not empty. Basically this function returns true if any of these functions return true: interval_overlaps, interval_overlapped_by, interval_covers, or interval_covered_by.

Syntax:

interval_overlapping(interval1, interval2)

This functions check whether two intervals share any points with each other.
Arguments:
- interval1, interval2: two intervals to be compared
Return Value:
- a boolean value. Specifically, interval_overlapping(interval1, interval2) is true if
  
  interval1.start < interval2.end AND interval1.end > interval2.start
- missing if the argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-interval input value will cause a type error.

Examples:

{
  "overlapping1": interval_overlapping(interval(date("2000-01-01"), date("2005-01-01")),
                                       interval(date("2004-05-01"), date("2012-09-09"))),
  "overlapping2": interval_overlapping(interval(date("2006-08-01"), date("2007-03-01")),
                                       interval(date("2004-09-10"), date("2006-12-31")))
};

The expected result is:

{ "overlapping1": true, "overlapping2": true }

interval_meets, interval_met_by

Syntax:

interval_meets(interval1, interval2)
interval_met_by(interval1, interval2)

These two functions check whether an interval meets with another interval.
Arguments:
- interval1, interval2: two intervals to be compared
Return Value:
- a boolean value. Specifically, interval_meets(interval1, interval2) is true if and only if interval1.end = interval2.start, and interval_met_by(interval1, interval2) is true if and only if interval1.start = interval2.end. If any of the two inputs is null, null is returned.
- missing if the argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-interval input value will cause a type error.

Examples:

{
  "meets": interval_meets(interval(date("2000-01-01"), date("2005-01-01")),
                          interval(date("2005-01-01"), date("2012-09-09"))),
  "metby": interval_met_by(interval(date("2006-08-01"), date("2007-03-01")),
                           interval(date("2004-09-10"), date("2006-08-01")))
};

The expected result is:
```
{ "meets": true, "metby": true }
```

interval_starts, interval_started_by

Syntax:

interval_starts(interval1, interval2)
interval_started_by(interval1, interval2)

These two functions check whether one interval starts with the other interval.
Arguments:
- interval1, interval2: two intervals to be compared
Return Value:
- a boolean value. Specifically, interval_starts(interval1, interval2) returns true if and only if
  interval1.start = interval2.start AND interval1.end <= interval2.end
interval_started_by(interval1, interval2) returns true if and only if
```
interval1.start = interval2.start
AND interval2.end <= interval1.end
```
- missing if the argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-interval input value will cause a type error.

Examples:

{
  "interval_starts": interval_starts(interval(date("2000-01-01"), date("2005-01-01")),
                                     interval(date("2000-01-01"), date("2012-09-09"))),
  "interval_started_by": interval_started_by(interval(date("2006-08-01"), date("2007-03-01")),
                                             interval(date("2006-08-01"), date("2006-08-02")))
};

The expected result is:

{ "interval_starts": true, "interval_started_by": true }

interval_ends, interval_ended_by

Syntax:

interval_ends(interval1, interval2)
interval_ended_by(interval1, interval2)

These two functions check whether one interval ends with the other interval.
Arguments:
- interval1, interval2: two intervals to be compared
Return Value:
- a boolean value. Specifically, interval_ends(interval1, interval2) returns true if and only if
  
  interval1.end = interval2.end AND interval1.start >= interval2.start
  
  interval_ended_by(interval1, interval2) returns true if and only if
  
  interval2.end = interval1.end AND interval2.start >= interval1.start
- missing if the argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-interval input value will cause a type error.

Examples:

{
  "interval_ends": interval_ends(interval(date("2000-01-01"), date("2005-01-01")),
                                 interval(date("1998-01-01"), date("2005-01-01"))),
  "interval_ended_by": interval_ended_by(interval(date("2006-08-01"), date("2007-03-01")),
                                         interval(date("2006-09-10"), date("2007-03-01")))
};

The expected result is:

{ "interval_ends": true, "interval_ended_by": true }

Object Functions

get_object_fields

Syntax:
```
get_object_fields(input_object)
```
Access the object field names, type and open status for a given object.
Arguments:
- input_object : a object value.
Return Value:
- an array of object values that include the field_name string, field_type string, is_open boolean (used for debug purposes only: true if field is open and false otherwise), and optional nested orderedList for the values of a nested object,
- missing if the argument is a missing value,
- null if the argument is a null value,
- any other non-object input value will cause a type error.

Example:

get_object_fields(
                  {
                    "id": 1,
                    "project": "AsterixDB",
                    "address": {"city": "Irvine", "state": "CA"},
                    "related": ["Hivestrix", "Preglix", "Apache VXQuery"]
                  }
                 );

The expected result is:

[
  { "field-name": "id", "field-type": "INT64", "is-open": false },
  { "field-name": "project", "field-type": "STRING", "is-open": false },
  { "field-name": "address", "field-type": "RECORD", "is-open": false,
    "nested": [
                { "field-name": "city", "field-type": "STRING", "is-open": false },
                { "field-name": "state", "field-type": "STRING", "is-open": false }
              ]
  },
  { "field-name":
        "related",
        "field-type": "ORDEREDLIST",
        "is-open": false,
        "list": [
                  { "field-type": "STRING" },
                  { "field-type": "STRING" },
                  { "field-type": "STRING" }
                ]
  }
]

]

get_object_field_value

Syntax:

get_object_field_value(input_object, string)

Access the field name given in the string_expression from the object_expression.
Arguments:
- input_object : a object value.
- string : a string representing the top level field name.
Return Value:
- an any value saved in the designated field of the object,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- a type error will be raised if:
  - the first argument is any other non-object value,
  - or, the second argument is any other non-string value.

Example:

get_object_field_value({
                         "id": 1,
                         "project": "AsterixDB",
                         "address": {"city": "Irvine", "state": "CA"},
                         "related": ["Hivestrix", "Preglix", "Apache VXQuery"]
                        },
                        "project"
                       );

The expected result is:
```
"AsterixDB"
```

object_remove_fields

Syntax:

object_remove_fields(input_object, field_names)

Remove indicated fields from a object given a list of field names.
Arguments:
- input_object: a object value.
- field_names: an array of strings and/or array of array of strings.
Return Value:
- a new object value without the fields listed in the second argument,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- a type error will be raised if:
  - the first argument is any other non-object value,
  - or, the second argument is any other non-array value or recursively contains non-string items.

Example:

object_remove_fields(
                       {
                         "id":1,
                         "project":"AsterixDB",
                         "address":{"city":"Irvine", "state":"CA"},
                         "related":["Hivestrix", "Preglix", "Apache VXQuery"]
                       },
                       [["address", "city"], "related"]
                     );

The expected result is:

{
  "id":1,
  "project":"AsterixDB",
  "address":{ "state": "CA" }
}

object_add_fields

Syntax:

object_add_fields(input_object, fields)

Add fields to a object given a list of field names.
Arguments:
- input_object : a object value.
- fields: an array of field descriptor objects where each object has field_name and field_value.
Return Value:
- a new object value with the new fields included,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- a type error will be raised if:
  - the first argument is any other non-object value,
  - the second argument is any other non-array value, or contains non-object items.

Example:

object_add_fields(
                   {
                     "id":1,
                     "project":"AsterixDB",
                     "address":{"city":"Irvine", "state":"CA"},
                     "related":["Hivestrix", "Preglix", "Apache VXQuery"]
                    },
                    [{"field-name":"employment_location", "field-value":create_point(30.0,70.0)}]
                  );

The expected result is:

{
   "id":1,
   "project":"AsterixDB",
   "address":{"city":"Irvine", "state":"CA"},
   "related":["Hivestrix", "Preglix", "Apache VXQuery"]
   "employment_location": point("30.0,70.0")
 }

object_merge

Syntax:
```
object_merge(object1, object2)
```
Merge two different objects into a new object.
Arguments:
- object1 : a object value.
- object2 : a object value.
Return Value:
- a new object value with fields from both input objects. If a field’s names in both objects are the same, an exception is issued,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- any other non-object input value will cause a type error.

Example:

object_merge(
              {
                "id":1,
                "project":"AsterixDB",
                "address":{"city":"Irvine", "state":"CA"},
                "related":["Hivestrix", "Preglix", "Apache VXQuery"]
              },
              {
                "user_id": 22,
                "employer": "UC Irvine",
                "employment_type": "visitor"
              }
            );

The expected result is:

{
  "employment_type": "visitor",
  "address": {
    "city": "Irvine",
    "state": "CA"
  },
  "related": [
    "Hivestrix",
    "Preglix",
    "Apache VXQuery"
  ],
  "user_id": 22,
  "project": "AsterixDB",
  "employer": "UC Irvine",
  "id": 1
}

object_length

Syntax:
```
object_length(input_object)
```
Returns number of top-level fields in the given object
Arguments:
- input_object : an object value.
Return Value:
- an integer that represents the number of top-level fields in the given object,
- missing if the argument is a missing value,
- null if the argument is a null value or any other non-object value

Example:

object_length(
               {
                 "id": 1,
                 "project": "AsterixDB",
                 "address": {"city": "Irvine", "state": "CA"},
               }
             );

The expected result is:
```
3
```

object_names

Syntax:
```
object_names(input_object)
```
Returns names of top-level fields in the given object
Arguments:
- input_object : an object value.
Return Value:
- an array with top-level field names of the given object,
- missing if the argument is a missing value,
- null if the argument is a null value or any other non-object value

Example:

object_names(
               {
                 "id": 1,
                 "project": "AsterixDB",
                 "address": {"city": "Irvine", "state": "CA"},
               }
             );

The expected result is:
```
[ "id", "project", "address" ]
```

Aggregate Functions (Array Functions)

This section contains detailed descriptions of each SQL++ aggregate function (i.e., array function). Note that as described in the SQL++ query reference documentation, standard SQL aggregate functions (e.g., MIN, MAX, SUM, COUNT, and AVG) are not real functions in SQL++ but just syntactic sugars over corresponding SQL++ builtin aggregate functions (e.g., ARRAY_MIN, ARRAY_MAX, ARRAY_SUM, ARRAY_COUNT, and ARRAY_AVG).

array_count

Syntax:
```
array_count(collection)
```
Gets the number of non-null and non-missing items in the given collection.
Arguments:
- collection could be:
  - an array or multiset to be counted,
  - or, a null value,
  - or, a missing value.
Return Value:
- a bigint value representing the number of non-null and non-missing items in the given collection,
- null is returned if the input is null or missing,
- any other non-array and non-multiset input value will cause an error.

Example:

array_count( ['hello', 'world', 1, 2, 3, null, missing] );

The expected result is:
```
5
```

array_avg

Syntax:
```
array_avg(num_collection)
```
Gets the average value of the non-null and non-missing numeric items in the given collection.
Arguments:
- num_collection could be:
  - an array or multiset containing numeric values, nulls or missings,
  - or, a null value,
  - or, a missing value.
Return Value:
- a double value representing the average of the non-null and non-missing numbers in the given collection,
- null is returned if the input is null or missing,
- null is returned if the given collection does not contain any non-null and non-missing items,
- any other non-array and non-multiset input value will cause a type error,
- any other non-numeric value in the input collection will cause a type error.
Example:
```
array_avg( [1.2, 2.3, 3.4, 0, null] );
```
The expected result is:
```
1.725
```

array_sum

Syntax:
```
array_sum(num_collection)
```
Gets the sum of non-null and non-missing items in the given collection.
Arguments:
- num_collection could be:
  - an array or multiset containing numeric values, nulls or missings,
  - or, a null value,
  - or, a missing value.
Return Value:
- the sum of the non-null and non-missing numbers in the given collection. The returning type is decided by the item type with the highest order in the numeric type promotion order (tinyint-> smallint->integer->bigint->float->double) among items.
- null is returned if the input is null or missing,
- null is returned if the given collection does not contain any non-null and non-missing items,
- any other non-array and non-multiset input value will cause a type error,
- any other non-numeric value in the input collection will cause a type error.

Example:

array_sum( [1.2, 2.3, 3.4, 0, null, missing] );

The expected result is:
```
6.9
```

array_sql_min

Syntax:
```
array_min(num_collection)
```
Gets the min value of non-null and non-missing comparable items in the given collection.
Arguments:
- num_collection could be:
  - an array or multiset,
  - or, a null value,
  - or, a missing value.
Return Value:
- the min value of non-null and non-missing values in the given collection. The returning type is decided by the item type with the highest order in the type promotion order (tinyint-> smallint->integer->bigint->float->double) among numeric items.
- null is returned if the input is null or missing,
- null is returned if the given collection does not contain any non-null and non-missing items,
- multiple incomparable items in the input array or multiset will cause a type error,
- any other non-array and non-multiset input value will cause a type error.

Example:

array_min( [1.2, 2.3, 3.4, 0, null, missing] );

The expected result is:
```
0.0
```

array_max

Syntax:
```
array_max(num_collection)
```
Gets the max value of the non-null and non-missing comparable items in the given collection.
Arguments:
- num_collection could be:
  - an array or multiset,
  - or, a null value,
  - or, a missing value.
Return Value:
- the max value of non-null and non-missing numbers in the given collection. The returning type is decided by the item type with the highest order in the type promotion order (tinyint-> smallint->integer->bigint->float->double) among numeric items.
- null is returned if the input is null or missing,
- null is returned if the given collection does not contain any non-null and non-missing items,
- multiple incomparable items in the input array or multiset will cause a type error,
- any other non-array and non-multiset input value will cause a type error.

Example:

array_max( [1.2, 2.3, 3.4, 0, null, missing] );

The expected result is:
```
3.4
```

coll_count

Syntax:
```
coll_count(collection)
```
Gets the number of items in the given collection.
Arguments:
- collection could be:
  - an array or multiset containing the items to be counted,
  - or a null value,
  - or a missing value.
Return Value:
- a bigint value representing the number of items in the given collection,
- null is returned if the input is null or missing.
Example:
```
coll_count( [1, 2, null, missing] );
```
The expected result is:
```
4
```

coll_avg

Syntax:
```
coll_avg(num_collection)
```
Gets the average value of the numeric items in the given collection.
Arguments:
- num_collection could be:
  - an array or multiset containing numeric values, nulls or missings,
  - or, a null value,
  - or, a missing value.
Return Value:
- a double value representing the average of the numbers in the given collection,
- null is returned if the input is null or missing,
- null is returned if there is a null or missing in the input collection,
- any other non-numeric value in the input collection will cause a type error.
Example:
```
coll_avg( [100, 200, 300] );
```
The expected result is:
```
[ 200.0 ]
```

coll_sum

Syntax:
```
coll_sum(num_collection)
```
Gets the sum of the items in the given collection.
Arguments:
- num_collection could be:
  - an array or multiset containing numeric values, nulls or missings,
  - or, a null value,
  - or, a missing value.
Return Value:
- the sum of the numbers in the given collection. The returning type is decided by the item type with the highest order in the numeric type promotion order (tinyint-> smallint->integer->bigint->float->double) among items.
- null is returned if the input is null or missing,
- null is returned if there is a null or missing in the input collection,
- any other non-numeric value in the input collection will cause a type error.
Example:
```
coll_sum( [100, 200, 300] );
```
The expected result is:
```
600
```

array_min

Syntax:
```
coll_min(num_collection)
```
Gets the min value of comparable items in the given collection.
Arguments:
- num_collection could be:
  - an array or multiset,
  - or, a null value,
  - or, a missing value.
Return Value:
- the min value of the given collection. The returning type is decided by the item type with the highest order in the type promotion order (tinyint-> smallint->integer->bigint->float->double) among numeric items.
- null is returned if the input is null or missing,
- null is returned if there is a null or missing in the input collection,
- multiple incomparable items in the input array or multiset will cause a type error,
- any other non-array and non-multiset input value will cause a type error.
Example:
```
coll_min( [10.2, 100, 5] );
```
The expected result is:
```
5.0
```

array_max

Syntax:
```
coll_max(num_collection)
```
Gets the max value of numeric items in the given collection.
Arguments:
- num_collection could be:
  - an array or multiset,
  - or, a null value,
  - or, a missing value.
Return Value:
- The max value of the given collection. The returning type is decided by the item type with the highest order in the type promotion order (tinyint-> smallint->integer->bigint->float->double) among numeric items.
- null is returned if the input is null or missing,
- null is returned if there is a null or missing in the input collection,
- multiple incomparable items in the input array or multiset will cause a type error,
- any other non-array and non-multiset input value will cause a type error.
Example:
```
coll_max( [10.2, 100, 5] );
```
The expected result is:
```
100.0
```

Comparison Functions

greatest

Syntax:

greatest(numeric_value1, numeric_value2, ...)

Computes the greatest value among arguments.
Arguments:
- numeric_value1: a tinyint/smallint/integer/bigint/float/double value,
- numeric_value2: a tinyint/smallint/integer/bigint/float/double value,
- ….
Return Value:
- the greatest values among arguments. The returning type is decided by the item type with the highest order in the numeric type promotion order (tinyint-> smallint->integer->bigint->float->double) among items.
- null if any argument is a missing value or null value,
- any other non-numeric input value will cause a type error.

Example:

{ "v1": greatest(1, 2, 3), "v2": greatest(float("0.5"), double("-0.5"), 5000) };

The expected result is:
```
{ "v1": 3, "v2": 5000.0 }
```

least

Syntax:

least(numeric_value1, numeric_value2, ...)

Computes the least value among arguments.
Arguments:
- numeric_value1: a tinyint/smallint/integer/bigint/float/double value,
- numeric_value2: a tinyint/smallint/integer/bigint/float/double value,
- ….
Return Value:
- the least values among arguments. The returning type is decided by the item type with the highest order in the numeric type promotion order (tinyint-> smallint->integer->bigint->float->double) among items.
- null if any argument is a missing value or null value,
- any other non-numeric input value will cause a type error.

Example:

{ "v1": least(1, 2, 3), "v2": least(float("0.5"), double("-0.5"), 5000) };

The expected result is:
```
{ "v1": 1, "v2": -0.5 }
```

Type Functions

is_array

Syntax:
```
is_array(expr)
```
Checks whether the given expression is evaluated to be an array value.
Arguments:
- expr : an expression (any type is allowed).
Return Value:
- a boolean on whether the argument is an array value or not,
- a missing if the argument is a missing value,
- a null if the argument is a null value.

Example:

{
  "a": is_array(true),
  "b": is_array(false),
  "c": isarray(null),
  "d": isarray(missing),
  "e": isarray("d"),
  "f": isarray(4.0),
  "g": isarray(5),
  "h": isarray(["1", 2]),
  "i": isarray({"a":1})
};

The expected result is:

{ "a": false, "b": false, "c": null, "e": false, "f": false, "g": false, "h": true, "i": false }

The function has an alias isarray.

is_atomic (is_atom)

Syntax:
```
is_atomic(expr)
```
Checks whether the given expression is evaluated to be a value of a primitive type.
Arguments:
- expr : an expression (any type is allowed).
Return Value:
- a boolean on whether the argument is a primitive type or not,
- a missing if the argument is a missing value,
- a null if the argument is a null value.

Example:

{
  "a": is_atomic(true),
  "b": is_atomic(false),
  "c": isatomic(null),
  "d": isatomic(missing),
  "e": isatomic("d"),
  "f": isatom(4.0),
  "g": isatom(5),
  "h": isatom(["1", 2]),
  "i": isatom({"a":1})
};

The expected result is:

{ "a": true, "b": true, "c": null, "e": true, "f": true, "g": true, "h": false, "i": false }

The function has three aliases: isatomic, is_atom, and isatom.

is_boolean (is_bool)

Syntax:
```
is_boolean(expr)
```
Checks whether the given expression is evaluated to be a boolean value.
Arguments:
- expr : an expression (any type is allowed).
Return Value:
- a boolean on whether the argument is a boolean value or not,
- a missing if the argument is a missing value,
- a null if the argument is a null value.

Example:

{
  "a": isboolean(true),
  "b": isboolean(false),
  "c": is_boolean(null),
  "d": is_boolean(missing),
  "e": isbool("d"),
  "f": isbool(4.0),
  "g": isbool(5),
  "h": isbool(["1", 2]),
  "i": isbool({"a":1})
};

The expected result is:

{ "a": true, "b": true, "c": null, "e": false, "f": false, "g": false, "h": false, "i": false }

The function has three aliases: isboolean, is_bool, and isbool.

is_number (is_num)

Syntax:
```
is_number(expr)
```
Checks whether the given expression is evaluated to be a numeric value.
Arguments:
- expr : an expression (any type is allowed).
Return Value:
- a boolean on whether the argument is a smallint/tinyint/integer/bigint/float/double value or not,
- a missing if the argument is a missing value,
- a null if the argument is a null value.

Example:

{
  "a": is_number(true),
  "b": is_number(false),
  "c": isnumber(null),
  "d": isnumber(missing),
  "e": isnumber("d"),
  "f": isnum(4.0),
  "g": isnum(5),
  "h": isnum(["1", 2]),
  "i": isnum({"a":1})
};

The expected result is:

{ "a": false, "b": false, "c": null, "e": false, "f": true, "g": true, "h": false, "i": false }

The function has three aliases: isnumber, is_num, and isnum.

is_object (is_obj)

Syntax:
```
is_object(expr)
```
Checks whether the given expression is evaluated to be a object value.
Arguments:
- expr : an expression (any type is allowed).
Return Value:
- a boolean on whether the argument is a object value or not,
- a missing if the argument is a missing value,
- a null if the argument is a null value.

Example:

{
  "a": is_object(true),
  "b": is_object(false),
  "c": isobject(null),
  "d": isobject(missing),
  "e": isobj("d"),
  "f": isobj(4.0),
  "g": isobj(5),
  "h": isobj(["1", 2]),
  "i": isobj({"a":1})
};

The expected result is:

{ “a”: false, “b”: false, “c”: null, “e”: false, “f”: false, “g”: false, “h”: false, “i”: true }

The function has three aliases: isobject, is_obj, and isobj.

is_string (is_str)

Syntax:
```
is_string(expr)
```
Checks whether the given expression is evaluated to be a string value.
Arguments:
- expr : an expression (any type is allowed).
Return Value:
- a boolean on whether the argument is a string value or not,
- a missing if the argument is a missing value,
- a null if the argument is a null value.

Example:

{
  "a": is_string(true),
  "b": isstring(false),
  "c": isstring(null),
  "d": isstr(missing),
  "e": isstr("d"),
  "f": isstr(4.0),
  "g": isstr(5),
  "h": isstr(["1", 2]),
  "i": isstr({"a":1})
};

The expected result is:

{ "a": false, "b": false, "c": null, "e": true, "f": false, "g": false, "h": false, "i": false }

The function has three aliases: isstring, is_str, and isstr.

is_null

Syntax:
```
is_null(expr)
```
Checks whether the given expression is evaluated to be a null value.
Arguments:
- expr : an expression (any type is allowed).
Return Value:
- a boolean on whether the variable is a null or not,
- a missing if the input is missing.

Example:

{ "v1": is_null(null), "v2": is_null(1), "v3": is_null(missing) };

The expected result is:
```
{ "v1": true, "v2": false }
```

The function has an alias isnull.

is_missing

Syntax:
```
is_missing(expr)
```
Checks whether the given expression is evaluated to be a missing value.
Arguments:
- expr : an expression (any type is allowed).
Return Value:
- a boolean on whether the variable is a missing or not.

Example:

{ "v1": is_missing(null), "v2": is_missing(1), "v3": is_missing(missing) };

The expected result is:

{ "v1": false, "v2": false, "v3": true }

The function has an alias ismissing.

is_unknown

Syntax:
```
is_unknown(expr)
```
Checks whether the given variable is a null value or a missing value.
Arguments:
- expr : an expression (any type is allowed).
Return Value:
- a boolean on whether the variable is a null/``missingvalue (true) or not (false`).

Example:

{ "v1": is_unknown(null), "v2": is_unknown(1), "v3": is_unknown(missing) };

The expected result is:

{ "v1": true, "v2": false, "v3": true }

The function has an alias isunknown.

to_array

Syntax:
```
to_array(expr)
```
Converts input value to an array value
Arguments:
- expr : an expression
Return Value:
- if the argument is missing then missing is returned
- if the argument is null then null is returned
- if the argument is of array type then it is returned as is
- if the argument is of multiset type then it is returned as an array with elements in an undefined order
- otherwise an array containing the input expression as its single item is returned

Example:

{
  "v1": to_array("asterix"),
  "v2": to_array(["asterix"]),
};

The expected result is:

{ "v1": ["asterix"], "v2": ["asterix"] }

The function has an alias toarray.

to_atomic (to_atom)

Syntax:
```
to_atomic(expr)
```
Converts input value to a primitive value
Arguments:
- expr : an expression
Return Value:
- if the argument is missing then missing is returned
- if the argument is null then null is returned
- if the argument is of primitive type then it is returned as is
- if the argument is of array or multiset type and has only one element then the result of invoking to_atomic() on that element is returned
- if the argument is of object type and has only one field then the result of invoking to_atomic() on the value of that field is returned
- otherwise null is returned

Example:

{
  "v1": to_atomic("asterix"),
  "v2": to_atomic(["asterix"]),
  "v3": to_atomic([0, 1]),
  "v4": to_atomic({"value": "asterix"}),
  "v5": to_number({"x": 1, "y": 2})
};

The expected result is:

{ "v1": "asterix", "v2": "asterix", "v3": null, "v4": "asterix", "v5": null }

The function has three aliases: toatomic, to_atom, and toatom.

to_boolean (to_bool)

Syntax:
```
to_boolean(expr)
```
Converts input value to a boolean value
Arguments:
- expr : an expression
Return Value:
- if the argument is missing then missing is returned
- if the argument is null then null is returned
- if the argument is of boolean type then it is returned as is
- if the argument is of numeric type then false is returned if it is 0 or NaN, otherwise true
- if the argument is of string type then false is returned if it’s empty, otherwise true
- if the argument is of array or multiset type then false is returned if it’s size is 0, otherwise true
- if the argument is of object type then false is returned if it has no fields, otherwise true
- type error is raised for all other input types

Example:

{
  "v1": to_boolean(0),
  "v2": to_boolean(1),
  "v3": to_boolean(""),
  "v4": to_boolean("asterix")
};

The expected result is:

{ "v1": false, "v2": true, "v3": false, "v4": true }

The function has three aliases: toboolean, to_bool, and tobool.

to_bigint

Syntax:
```
to_bigint(expr)
```
Converts input value to an integer value
Arguments:
- expr : an expression
Return Value:
- if the argument is missing then missing is returned
- if the argument is null then null is returned
- if the argument is of boolean type then 1 is returned if it is true, 0 if it is false
- if the argument is of numeric integer type then it is returned as the same value of bigint type
- if the argument is of numeric float/double type then it is converted to bigint type
- if the argument is of string type and can be parsed as integer then that integer value is returned, otherwise null is returned
- if the argument is of array/multiset/object type then null is returned
- type error is raised for all other input types

Example:

{
  "v1": to_bigint(false),
  "v2": to_bigint(true),
  "v3": to_bigint(10),
  "v4": to_bigint(float("1e100")),
  "v5": to_bigint(double("1e1000")),
  "v6": to_bigint("20")
};

The expected result is:

{ "v1": 0, "v2": 1, "v3": 10, "v4": 9223372036854775807, "v5": 9223372036854775807, "v6": 20 }

The function has an alias tobigint.

to_double

Syntax:
```
to_double(expr)
```
Converts input value to a double value
Arguments:
- expr : an expression
Return Value:
- if the argument is missing then missing is returned
- if the argument is null then null is returned
- if the argument is of boolean type then 1.0 is returned if it is true, 0.0 if it is false
- if the argument is of numeric type then it is returned as the value of double type
- if the argument is of string type and can be parsed as double then that double value is returned, otherwise null is returned
- if the argument is of array/multiset/object type then null is returned
- type error is raised for all other input types

Example:

{
  "v1": to_double(false),
  "v2": to_double(true),
  "v3": to_double(10),
  "v4": to_double(11.5),
  "v5": to_double("12.5")
};

The expected result is:

{ "v1": 0.0, "v2": 1.0, "v3": 10.0, "v4": 11.5, "v5": 12.5 }

The function has an alias todouble.

to_number (to_num)

Syntax:
```
to_number(expr)
```
Converts input value to a numeric value
Arguments:
- expr : an expression
Return Value:
- if the argument is missing then missing is returned
- if the argument is null then null is returned
- if the argument is of numeric type then it is returned as is
- if the argument is of boolean type then 1 is returned if it is true, 0 if it is false
- if the argument is of string type and can be parsed as bigint then that bigint value is returned, otherwise if it can be parsed as double then that double value is returned, otherwise null is returned
- if the argument is of array/multiset/object type then null is returned
- type error is raised for all other input types

Example:

{
  "v1": to_number(false),
  "v2": to_number(true),
  "v3": to_number(10),
  "v4": to_number(11.5),
  "v5": to_number("12.5")
};

The expected result is:

{ "v1": 0, "v2": 1, "v3": 10, "v4": 11.5, "v5": 12.5 }

The function has three aliases: tonumber, to_num, and tonum.

to_object (to_obj)

Syntax:
```
to_object(expr)
```
Converts input value to an object value
Arguments:
- expr : an expression
Return Value:
- if the argument is missing then missing is returned
- if the argument is null then null is returned
- if the argument is of object type then it is returned as is
- otherwise an empty object is returned

Example:

{
  "v1": to_object({"value": "asterix"}),
  "v2": to_object("asterix")
};

The expected result is:

{ "v1": {"value": "asterix"}, "v2": {} }

The function has three aliases: toobject, to_obj, and toobj.

to_string (to_str)

Syntax:
```
to_string(expr)
```
Converts input value to a string value
Arguments:
- expr : an expression
Return Value:
- if the argument is missing then missing is returned
- if the argument is null then null is returned
- if the argument is of boolean type then "true" is returned if it is true, "false" if it is false
- if the argument is of numeric type then its string representation is returned
- if the argument is of string type then it is returned as is
- if the argument is of array/multiset/object type then null is returned
- type error is raised for all other input types

Example:

{
  "v1": to_string(false),
  "v2": to_string(true),
  "v3": to_string(10),
  "v4": to_string(11.5),
  "v5": to_string("asterix")
};

The expected result is:

{ "v1": "false", "v2": "true", "v3": "10", "v4": "11.5", "v5": "asterix" }

The function has three aliases: tostring, to_str, and tostr.

Conditional Functions

if_null (ifnull)

Syntax:

if_null(expression1, expression2, ... expressionN)

Finds first argument which value is not null and returns that value
Arguments:
- expressionI : an expression (any type is allowed).
Return Value:
- a null if all arguments evaluate to null or no arguments specified
- a value of the first non-null argument otherwise

Example:

{
    "a": if_null(),
    "b": if_null(null),
    "c": if_null(null, "asterixdb"),
    "d": is_missing(if_null(missing))
};

The expected result is:

{ "a": null, "b": null, "c": "asterixdb", "d": true }

The function has an alias ifnull.

if_missing (ifmissing)

Syntax:

if_missing(expression1, expression2, ... expressionN)

Finds first argument which value is not missing and returns that value
Arguments:
- expressionI : an expression (any type is allowed).
Return Value:
- a null if all arguments evaluate to missing or no arguments specified
- a value of the first non-missing argument otherwise

Example:

{
    "a": if_missing(),
    "b": if_missing(missing),
    "c": if_missing(missing, "asterixdb"),
    "d": if_missing(null, "asterixdb")
};

The expected result is:

{ "a": null, "b": null, "c": "asterixdb", "d": null }

The function has an alias ifmissing.

if_missing_or_null (ifmissingornull)

Syntax:

if_missing_or_null(expression1, expression2, ... expressionN)

Finds first argument which value is not null or missing and returns that value
Arguments:
- expressionI : an expression (any type is allowed).
Return Value:
- a null if all arguments evaluate to either null or missing, or no arguments specified
- a value of the first non-null, non-missing argument otherwise

Example:

{
    "a": if_missing_or_null(),
    "b": if_missing_or_null(null, missing),
    "c": if_missing_or_null(null, missing, "asterixdb")
};

The expected result is:

{ "a": null, "b": null, "c": "asterixdb" }

The function has an alias ifmissingornull.

if_inf (ifinf)

Syntax:

if_inf(expression1, expression2, ... expressionN)

Finds first argument which is a non-infinite (INF or-INF) number
Arguments:
- expressionI : an expression (any type is allowed).
Return Value:
- a missing if missing argument was encountered before the first non-infinite number argument
- a null if null argument or any other non-number argument was encountered before the first non-infinite number argument
- the first non-infinite number argument otherwise

Example:

{
    "a": is_null(if_inf(null)),
    "b": is_missing(if_inf(missing)),
    "c": is_null(if_inf(double("INF"))),
    "d": if_inf(1, null, missing) ],
    "e": is_null(if_inf(null, missing, 1)) ],
    "f": is_missing(if_inf(missing, null, 1)) ],
    "g": if_inf(float("INF"), 1) ],
    "h": to_string(if_inf(float("INF"), double("NaN"), 1)) ]
};

The expected result is:

{ "a": true, "b": true, "c": true, "d": 1, "e": true, "f": true, "g": 1, "h": "NaN" }

The function has an alias ifinf.

if_nan (ifnan)

Syntax:

if_nan(expression1, expression2, ... expressionN)

Finds first argument which is a non-NaN number
Arguments:
- expressionI : an expression (any type is allowed).
Return Value:
- a missing if missing argument was encountered before the first non-NaN number argument
- a null if null argument or any other non-number argument was encountered before the first non-NaN number argument
- the first non-NaN number argument otherwise

Example:

{
    "a": is_null(if_nan(null)),
    "b": is_missing(if_nan(missing)),
    "c": is_null(if_nan(double("NaN"))),
    "d": if_nan(1, null, missing) ],
    "e": is_null(if_nan(null, missing, 1)) ],
    "f": is_missing(if_nan(missing, null, 1)) ],
    "g": if_nan(float("NaN"), 1) ],
    "h": to_string(if_nan(float("NaN"), double("INF"), 1)) ]
};

The expected result is:

{ "a": true, "b": true, "c": true, "d": 1, "e": true, "f": true, "g": 1, "h": "INF" }

The function has an alias ifnan.

if_nan_or_inf (ifnanorinf)

Syntax:

if_nan_or_inf(expression1, expression2, ... expressionN)

Finds first argument which is a non-infinite (INF or-INF) and non-NaN number
Arguments:
- expressionI : an expression (any type is allowed).
Return Value:
- a missing if missing argument was encountered before the first non-infinite and non-NaN number argument
- a null if null argument or any other non-number argument was encountered before the first non-infinite and non-NaN number argument
- the first non-infinite and non-NaN number argument otherwise

Example:

{
    "a": is_null(if_nan_or_inf(null)),
    "b": is_missing(if_nan_or_inf(missing)),
    "c": is_null(if_nan_or_inf(double("NaN"), double("INF"))),
    "d": if_nan_or_inf(1, null, missing) ],
    "e": is_null(if_nan_or_inf(null, missing, 1)) ],
    "f": is_missing(if_nan_or_inf(missing, null, 1)) ],
    "g": if_nan_or_inf(float("NaN"), float("INF"), 1) ],
};

The expected result is:

{ "a": true, "b": true, "c": true, "d": 1, "e": true, "f": true, "g": 1 }

The function has an alias ifnanorinf.

Miscellaneous Functions

uuid

Syntax:
```
uuid()
```
Generates a uuid.
Arguments:
- none
Return Value:
- a generated, random uuid.

len

Syntax:

len(array)
Returns the length of the array array.
Arguments:
- array : an array, multiset, null, or missing, represents the collection that needs to be checked.
Return Value:
- an integer that represents the length of input array or the size of the input multiset,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value.
Example:
```
len(["Hello", "World"])
```
The expected result is:
```
2
```

not

Syntax:
```
not(expr)
```
Inverts a boolean value
Arguments:
- expr : an expression
Return Value:
- a boolean, the inverse of expr,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value,
- other non-boolean argument value will cause a type error.

Example:

{ "v1": `not`(true), "v2": `not`(false), "v3": `not`(null), "v4": `not`(missing) };

The expected result is:

{ "v1": false, "v2": true, "v3": null }

range

Syntax:

range(start_numeric_value, end_numeric_value)

Generates a series of bigint values based start the start_numeric_value until the end_numeric_value.
Arguments:
start_numeric_value: a tinyint/smallint/integer/bigint value representing the start value.
end_numeric_value: a tinyint/smallint/integer/bigint value representing the max final value.
Return Value:
- an array that starts with the integer value of start_numeric_value and ends with the integer value of end_numeric_value, where the value of each entry in the array is the integer successor of the value in the preceding entry.
Example:
```
range(0, 3);
```
The expected result is:
```
[ 0, 1, 2, 3 ]
```

switch_case

Syntax:

switch_case(
    condition,
    case1, case1_result,
    case2, case2_result,
    ...,
    default, default_result
)

Switches amongst a sequence of cases and returns the result of the first matching case. If no match is found, the result of the default case is returned.
Arguments:
- condition: a variable (any type is allowed).
- caseI/default: a variable (any type is allowed).
- caseI/default_result: a variable (any type is allowed).
Return Value:
- caseI_result if condition matches caseI, otherwise default_result.

Example 1:

switch_case(
    "a",
    "a", 0,
    "x", 1,
    "y", 2,
    "z", 3
);

The expected result is:
```
0
```

Example 2:

switch_case(
    "a",
    "x", 1,
    "y", 2,
    "z", 3
);

The expected result is:
```
3
```

deep_equal

Syntax:
```
deep_equal(expr1, expr2)
```
Assess the equality between two expressions of any type (e.g., object, arrays, or multiset). Two objects are deeply equal iff both their types and values are equal.
Arguments:
- expr1 : an expression,
- expr2 : an expression.
Return Value:
- true or false depending on the data equality,
- missing if any argument is a missing value,
- null if any argument is a null value but no argument is a missing value.

Example:

deep_equal(
           {
             "id":1,
             "project":"AsterixDB",
             "address":{"city":"Irvine", "state":"CA"},
             "related":["Hivestrix", "Preglix", "Apache VXQuery"]
           },
           {
             "id":1,
             "project":"AsterixDB",
             "address":{"city":"San Diego", "state":"CA"},
             "related":["Hivestrix", "Preglix", "Apache VXQuery"]
           }
);

The expected result is:
```
false
```

Builtin Functions

abs

acos

asin

atan

atan2

ceil

cos

exp

floor

ln

log

power

round

sign

sin

sqrt

tan

trunc

round_half_to_even

concat

contains

ends_with

initcap (or title)

length

lower

ltrim

position

regexp_contains

regexp_like

regexp_position

regexp_replace

repeat

replace

rtrim

split

starts_with

substr

trim

upper

string_concat

string_join

string_to_codepoint

codepoint_to_string

substring_before

substring_after

parse_binary

print_binary

binary_length

sub_binary

binary_concat

create_point

create_line

create_rectangle

create_circle

create_polygon

get_x/get_y

get_points

get_center/get_radius

spatial_distance

spatial_area

spatial_intersect

spatial_cell

edit_distance

edit_distance_check

edit_distance_contains

similarity_jaccard

similarity_jaccard_check

word_tokens

get_year/get_month/get_day/get_hour/get_minute/get_second/get_millisecond

adjust_datetime_for_timezone

adjust_time_for_timezone

calendar_duration_from_datetime

get_year_month_duration/get_day_time_duration

months_from_year_month_duration/milliseconds_from_day_time_duration

duration_from_months/duration_from_ms

duration_from_interval

current_date

current_time

current_datetime