ArangoDB v3.9 reached End of Life (EOL) and is no longer supported.

This documentation is outdated. Please see the most recent version at **docs.arangodb.com**

# Array functions

AQL provides functions for higher-level array manipulation. Also see the
numeric functions for functions that work on number arrays.
If you want to concatenate the elements of an array equivalent to `join()`

in JavaScript, see CONCAT() and
CONCAT_SEPARATOR() in the string functions chapter.

Apart from that, AQL also offers several language constructs:

- simple array access of individual elements,
- array operators for array expansion and contraction, optionally with inline filter, limit and projection,
- array comparison operators to compare each element in an array to a value or the elements of another array,
- loop-based operations on arrays using FOR, SORT, LIMIT, as well as COLLECT for grouping, which also offers efficient aggregation.

## APPEND()

`APPEND(anyArray, values, unique) → newArray`

Add all elements of an array to another array. All values are added at the end of the array (right side).

It can also be used to append a single element to an array. It is not necessary to wrap it in an array (unless it is an array itself). You may also use PUSH() instead.

**anyArray**(array): array with elements of arbitrary type**values**(array|any): array, whose elements shall be added to`anyArray`

**unique**(bool,*optional*): if set to`true`

, all duplicate values are removed from the resulting array. If`values`

is an empty array or if either`anyArray`

or`values`

is`null`

, then the other input array is returned unmodified. The default is`false`

.- returns
**newArray**(array): the modified array

**Examples**

## CONTAINS_ARRAY()

This is an alias for POSITION().

## COUNT()

This is an alias for LENGTH().

## COUNT_DISTINCT()

`COUNT_DISTINCT(anyArray) → number`

Get the number of distinct elements in an array.

**anyArray**(array): array with elements of arbitrary type- returns
**number**: the number of distinct elements in*anyArray*.

**Examples**

`RETURN COUNT_DISTINCT([ 1, 2, 3 ])`

```
[
3
]
```

`RETURN COUNT_DISTINCT([ "yes", "no", "yes", "sauron", "no", "yes" ])`

```
[
3
]
```

## COUNT_UNIQUE()

This is an alias for COUNT_DISTINCT().

## FIRST()

`FIRST(anyArray) → firstElement`

Get the first element of an array. It is the same as `anyArray[0]`

.

**anyArray**(array): array with elements of arbitrary type- returns
**firstElement**(any|null): the first element of*anyArray*, or*null*if the array is empty.

**Examples**

`RETURN FIRST([ 1, 2, 3 ])`

```
[
1
]
```

RETURN FIRST([])

```
[
null
]
```

## FLATTEN()

`FLATTEN(anyArray, depth) → flatArray`

Turn an array of arrays into a flat array. All array elements in *array* will be
expanded in the result array. Non-array elements are added as they are. The function
will recurse into sub-arrays up to the specified depth. Duplicates will not be removed.

Also see array contraction.

**array**(array): array with elements of arbitrary type, including nested arrays**depth**(number,*optional*): flatten up to this many levels, the default is 1- returns
**flatArray**(array): a flattened array

**Examples**

To fully flatten the example array, use a *depth* of 2:

## INTERLEAVE()

Introduced in: v3.7.1

`INTERLEAVE(array1, array2, ... arrayN) → newArray`

Accepts an arbitrary number of arrays and produces a new array with the elements interleaved. It iterates over the input arrays in a round robin fashion, picks one element from each array per iteration, and combines them in that sequence into a result array. The input arrays can have different amounts of elements.

**arrays**(array,*repeatable*): an arbitrary number of arrays as multiple arguments (at least 2)- returns
**newArray**(array): the interleaved array

**Examples**

## INTERSECTION()

`INTERSECTION(array1, array2, ... arrayN) → newArray`

Return the intersection of all arrays specified. The result is an array of values that occur in all arguments.

Other set operations are UNION(), MINUS() and OUTERSECTION().

**arrays**(array,*repeatable*): an arbitrary number of arrays as multiple arguments (at least 2)- returns
**newArray**(array): a single array with only the elements, which exist in all provided arrays. The element order is random. Duplicates are removed.

**Examples**

`RETURN INTERSECTION( [2,4,6], [8,10,12], [14,16,18] )`

[ [] ]

## JACCARD()

Introduced in: v3.7.0

`JACCARD(array1, array2) → jaccardIndex`

Calculate the Jaccard index of two arrays.

This similarity measure is also known as *Intersection over Union* and could
be computed (less efficient and more verbose) as follows:

```
COUNT(a) == 0 && COUNT(b) == 0
? 1 // two empty sets have a similarity of 1 by definition
: COUNT(INTERSECTION(array1, array2)) / COUNT(UNION_DISTINCT(array1, array2))
```

**array1**(array): array with elements of arbitrary type**array2**(array): array with elements of arbitrary type- returns
**jaccardIndex**(number): calculated Jaccard index of the input arrays*array1*and*array2*

`RETURN JACCARD( [1,2,3,4], [3,4,5,6] )`

```
[
0.3333333333333333
]
```

`RETURN JACCARD( [1,1,2,2,2,3], [2,2,3,4] )`

```
[
0.5
]
```

`RETURN JACCARD( [1,2,3], [] )`

```
[
0
]
```

RETURN JACCARD( [], [] )

```
[
1
]
```

## LAST()

`LAST(anyArray) → lastElement`

Get the last element of an array. It is the same as `anyArray[-1]`

.

**anyArray**(array): array with elements of arbitrary type- returns
**lastElement**(any|null): the last element of*anyArray*or*null*if the array is empty.

**Example**

`RETURN LAST( [1,2,3,4,5] )`

```
[
5
]
```

## LENGTH()

`LENGTH(anyArray) → length`

Determine the number of elements in an array.

**anyArray**(array): array with elements of arbitrary type- returns
**length**(number): the number of array elements in*anyArray*.

*LENGTH()* can also determine the number of attribute keys
of an object / document, the amount of documents in a
collection and the character length of a string.

Input | Length |
---|---|

String | Number of Unicode characters |

Number | Number of Unicode characters that represent the number |

Array | Number of elements |

Object | Number of first level elements |

true | 1 |

false | 0 |

null | 0 |

**Examples**

`RETURN LENGTH( "🥑" )`

```
[
1
]
```

`RETURN LENGTH( 1234 )`

```
[
4
]
```

`RETURN LENGTH( [1,2,3,4,5,6,7] )`

```
[
7
]
```

`RETURN LENGTH( false )`

```
[
0
]
```

`RETURN LENGTH( {a:1, b:2, c:3, d:4, e:{f:5,g:6}} )`

```
[
5
]
```

## MINUS()

`MINUS(array1, array2, ... arrayN) → newArray`

Return the difference of all arrays specified.

Other set operations are UNION(), INTERSECTION() and OUTERSECTION().

**arrays**(array,*repeatable*): an arbitrary number of arrays as multiple arguments (at least 2)- returns
**newArray**(array): an array of values that occur in the first array, but not in any of the subsequent arrays. The order of the result array is undefined and should not be relied on. Duplicates will be removed.

**Example**

## NTH()

`NTH(anyArray, position) → nthElement`

Get the element of an array at a given position. It is the same as `anyArray[position]`

for positive positions, but does not support negative positions.

**anyArray**(array): array with elements of arbitrary type**position**(number): position of desired element in array, positions start at 0- returns
**nthElement**(any|null): the array element at the given*position*. If*position*is negative or beyond the upper bound of the array, then*null*will be returned.

**Examples**

`RETURN NTH( [ "foo", "bar", "baz" ], 2 )`

```
[
"baz"
]
```

`RETURN NTH( [ "foo", "bar", "baz" ], 3 )`

```
[
null
]
```

`RETURN NTH( [ "foo", "bar", "baz" ], -1 )`

```
[
null
]
```

## OUTERSECTION()

`OUTERSECTION(array1, array2, ... arrayN) → newArray`

Return the values that occur only once across all arrays specified.

Other set operations are UNION(), MINUS() and INTERSECTION().

**arrays**(array,*repeatable*): an arbitrary number of arrays as multiple arguments (at least 2)- returns
**newArray**(array): a single array with only the elements that exist only once across all provided arrays. The element order is random.

**Example**

## POP()

`POP(anyArray) → newArray`

Remove the last element of *array*.

To append an element (right side), see PUSH().

To remove the first element, see SHIFT().

To remove an element at an arbitrary position, see REMOVE_NTH().

**anyArray**(array): an array with elements of arbitrary type- returns
**newArray**(array):*anyArray*without the last element. If it’s already empty or has only a single element left, an empty array is returned.

**Examples**

`RETURN POP( [ 1 ] )`

[ [] ]

## POSITION()

`POSITION(anyArray, search, returnIndex) → position`

Return whether *search* is contained in *array*. Optionally return the position.

**anyArray**(array): the haystack, an array with elements of arbitrary type**search**(any): the needle, an element of arbitrary type**returnIndex**(bool,*optional*): if set to*true*, the position of the match is returned instead of a boolean. The default is*false*.- returns
**position**(bool|number):*true*if*search*is contained in*anyArray*,*false*otherwise. If*returnIndex*is enabled, the position of the match is returned (positions start at 0), or*-1*if it’s not found.

If you want to check if a value is in an array, you can alternatively use
the `IN`

operator, for example,
`3 IN [1, 2, 3]`

instead of `POSITION([1, 2, 3], 3)`

.

To determine if or at which position a string occurs in another string, see the CONTAINS() string function.

**Examples**

Test whether a value is contained in an array:

`RETURN POSITION( [2,4,6,8], 4 )`

```
[
true
]
```

Return the position of the match, i.e. the array index, or `-1`

if the value is
not contained in the array:

`RETURN POSITION( [2,4,6,8], 4, true )`

```
[
1
]
```

If you want to search a list of objects, you can use the
array expansion operator `[*]`

.
For example, you can get an attribute from each object using the operator, and
then determine the array index of the first match using the `POSITION()`

function:

```
LET arr = [ { value: "foo" }, { value: "bar" }, { value: "baz" }, { value: "bay"} ]
RETURN POSITION(arr[*].value, "baz", true)
```

```
[
2
]
```

If you are not interested in the actual position but only want to check for
existence, you may use the `IN`

operator instead of calling `POSITION()`

, like
`"baz" IN arr[*].value`

.

## PUSH()

`PUSH(anyArray, value, unique) → newArray`

Append *value* to *anyArray* (right side).

To remove the last element, see POP().

To prepend a value (left side), see UNSHIFT().

To append multiple elements, see APPEND().

**anyArray**(array): array with elements of arbitrary type**value**(any): an element of arbitrary type**unique**(bool): if set to*true*, then*value*is not added if already present in the array. The default is*false*.- returns
**newArray**(array):*anyArray*with*value*added at the end (right side)

Note: The *unique* flag only controls if *value* is added if it’s already present
in *anyArray*. Duplicate elements that already exist in *anyArray* will not be
removed. To make an array unique, use the UNIQUE() function.

**Examples**

## REMOVE_NTH()

`REMOVE_NTH(anyArray, position) → newArray`

Remove the element at *position* from the *anyArray*.

To remove the first element, see SHIFT().

To remove the last element, see POP().

**anyArray**(array): array with elements of arbitrary type**position**(number): the position of the element to remove. Positions start at 0. Negative positions are supported, with -1 being the last array element. If*position*is out of bounds, the array is returned unmodified.- returns
**newArray**(array):*anyArray*without the element at*position*

**Examples**

## REPLACE_NTH()

Introduced in: v3.7.0

`REPLACE_NTH(anyArray, position, replaceValue, defaultPaddingValue) → newArray`

Replace the element at *position* in *anyArray* with *replaceValue*.

**anyArray**(array): array with elements of arbitrary type**position**(number): the position of the element to replace. Positions start at 0. Negative positions are supported, with -1 being the last array element. If a negative*position*is out of bounds, then it is set to the first element (0)**replaceValue**the value to be inserted at*position***defaultPaddingValue**to be used for padding if*position*is two or more elements beyond the last element in*anyArray*- returns
**newArray**(array):*anyArray*with the element at*position*replaced by*replaceValue*, or appended to*anyArray*and possibly padded by*defaultPaddingValue*

It is allowed to specify a position beyond the upper array boundary:

*replaceValue*is appended if*position*is equal to the array length- if it is higher,
*defaultPaddingValue*is appended to*anyArray*as many times as needed to place*replaceValue*at*position* - if no
*defaultPaddingValue*is supplied in above case, then a query error is raised

**Examples**

Trying to access out of bounds, without providing a padding value will result in an error:

`arangosh> db._query('RETURN REPLACE_NTH( [ "a", "b", "c" ], 6 , "z")');`

```
[ArangoError 1542: AQL: invalid argument type in call to function 'REPLACE_NTH()' (while optimizing ast)]
```

## REMOVE_VALUE()

`REMOVE_VALUE(anyArray, value, limit) → newArray`

Remove all occurrences of *value* in *anyArray*. Optionally with a *limit*
to the number of removals.

**anyArray**(array): array with elements of arbitrary type**value**(any): an element of arbitrary type**limit**(number,*optional*): cap the number of removals to this value- returns
**newArray**(array):*anyArray*with*value*removed

**Examples**

## REMOVE_VALUES()

`REMOVE_VALUES(anyArray, values) → newArray`

Remove all occurrences of any of the *values* from *anyArray*.

**anyArray**(array): array with elements of arbitrary type**values**(array): an array with elements of arbitrary type, that shall be removed from*anyArray*- returns
**newArray**(array):*anyArray*with all individual*values*removed

**Example**

## REVERSE()

`REVERSE(anyArray) → reversedArray`

Return an array with its elements reversed.

**anyArray**(array): array with elements of arbitrary type- returns
**reversedArray**(array): a new array with all elements of*anyArray*in reversed order

**Example**

## SHIFT()

`SHIFT(anyArray) → newArray`

Remove the first element of *anyArray*.

To prepend an element (left side), see UNSHIFT().

To remove the last element, see POP().

To remove an element at an arbitrary position, see REMOVE_NTH().

**anyArray**(array): array with elements with arbitrary type- returns
**newArray**(array):*anyArray*without the left-most element. If*anyArray*is already empty or has only one element left, an empty array is returned.

**Examples**

`RETURN SHIFT( [ 1 ] )`

[ [] ]

## SLICE()

`SLICE(anyArray, start, length) → newArray`

Extract a slice of *anyArray*.

**anyArray**(array): array with elements of arbitrary type**start**(number): start extraction at this element. Positions start at 0. Negative values indicate positions from the end of the array.**length**(number,*optional*): extract up to*length*elements, or all elements from*start*up to*length*if negative (exclusive)- returns
**newArray**(array): the specified slice of*anyArray*. If*length*is not specified, all array elements starting at*start*will be returned.

**Examples**

`RETURN SLICE( [ 1, 2, 3, 4, 5 ], 0, 1 )`

```
[
[
1
]
]
```

## SORTED()

`SORTED(anyArray) → newArray`

Sort all elements in *anyArray*. The function will use the default comparison
order for AQL value types.

**anyArray**(array): array with elements of arbitrary type- returns
**newArray**(array):*anyArray*, with elements sorted

**Example**

## SORTED_UNIQUE()

`SORTED_UNIQUE(anyArray) → newArray`

Sort all elements in *anyArray*. The function will use the default comparison
order for AQL value types. Additionally, the values in the result array will
be made unique.

**anyArray**(array): array with elements of arbitrary type- returns
**newArray**(array):*anyArray*, with elements sorted and duplicates removed

**Example**

## UNION()

`UNION(array1, array2, ... arrayN) → newArray`

Return the union of all arrays specified.

Other set operations are MINUS(), INTERSECTION() and OUTERSECTION().

**arrays**(array,*repeatable*): an arbitrary number of arrays as multiple arguments (at least 2)- returns
**newArray**(array): all array elements combined in a single array, in any order

**Examples**

Note: No duplicates will be removed. In order to remove duplicates, please use
either UNION_DISTINCT()
or apply UNIQUE() on the
result of *UNION()*:

## UNION_DISTINCT()

`UNION_DISTINCT(array1, array2, ... arrayN) → newArray`

Return the union of distinct values of all arrays specified.

**arrays**(array,*repeatable*): an arbitrary number of arrays as multiple arguments (at least 2)- returns
**newArray**(array): the elements of all given arrays in a single array, without duplicates, in any order

**Example**

## UNIQUE()

`UNIQUE(anyArray) → newArray`

Return all unique elements in *anyArray*. To determine uniqueness, the
function will use the comparison order.

**anyArray**(array): array with elements of arbitrary type- returns
**newArray**(array):*anyArray*without duplicates, in any order

**Example**

## UNSHIFT()

`UNSHIFT(anyArray, value, unique) → newArray`

Prepend *value* to *anyArray* (left side).

To remove the first element, see SHIFT().

To append a value (right side), see PUSH().

**anyArray**(array): array with elements of arbitrary type**value**(any): an element of arbitrary type**unique**(bool): if set to*true*, then*value*is not added if already present in the array. The default is*false*.- returns
**newArray**(array):*anyArray*with*value*added at the start (left side)

Note: The *unique* flag only controls if *value* is added if it’s already present
in *anyArray*. Duplicate elements that already exist in *anyArray* will not be
removed. To make an array unique, use the UNIQUE() function.

**Examples**