Bash Arrays

From NovaOrdis Knowledge Base
Revision as of 19:15, 1 April 2024 by Ovidiu (talk | contribs) (→‎Get All Associative Array Values)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

External

Internal

Indexed Arrays

External

Indexed Arrays Overview

bash indexed arrays are 0-based and unidimensional. No explicit declaration is necessary if at least one element is initialized as described below.

Indexed arrays behave as local variables: an array declared in a function will be accessible to the declaring function and all functions invoked from the declaring function after declaration, but not to callers of the declaring function. In other words, indexed arrays are accessible only in the declaring scope and sub-scopes.

Indexed Arrays Declaration

Arrays do not need explicit declarations, they will be automatically declared upon the first assignment of any of their elements:

my_array_var[0]="something"

They can be explicitly declared, though:

declare -a my_array_var

Previously declared arrays can be listed with:

declare -a

Whether a specific array was declared can be checked as follows:

if declare -a | grep -q my_array_var; then
   # 'my_array_var' was declared
   ...
else
   # 'my_array_var' was not declared
   ...
fi

Negation also works:

if ! declare -a | grep -q my_array_var; then
   # 'my_array_var' was not declared
   ...
else
   # 'my_array_var' was declared   
   ...
fi

An indexed array is automatically declared if it is assigned using this syntax:

my_array_var=("a" "b" "c")

An array variable can be "undeclared" with the "unset" builtin.

unset my_array_var

Individual array elements can also be unset:

unset my_array_var[subscript] 

A subscript of "*" or "@" also removes the entire array.

Indexed Array Declaration Test

This sequence tests whether a variable was declared as an indexed array. The test does not tell whether the indexed array was actually allocated.

declare -a | grep -q "declare -a my_array" && \
 echo "'my_array' was declared as indexed array" || \
 echo "'my_array' was not declared as indexed array"

Indexed Array Allocation Test

This sequence tests whether an indexed array was actually allocated (can be empty) and associated with the variable.

[[ -v my_array[@] ]] && \
  echo "there is an array associated with 'my_array'" || \
  echo "there is NO array associated with 'my_array'"

The above detects allocation in both following cases:

my_array[0]=
my_array[0]="a"

Also see:

-v

Indexed Arrays Assignment

Assign Indexed Array Individual Elements

Individual elements are initialized by using ${var-name[index]}=value notation:

a[0]="something"

Entire Indexed Array

The (value ...) or ([index]=value ...) syntax can be used to initialize an entire array. The values must be separated by space:

a=("X" "Y" "Z")

If we want initialize only specific elements, we can use this syntax:

a=([0]="X" [4]="Y" [8]="Z")

Arrays initialized this way are automatically declared.

Append Array Elements

Use the += operator:

declare -a my_array
my_array+=(a)
my_array+=(b)
my_array+=(c)

Prepend Array Elements

declare -a my_array
my_array+=(b)

# this is how you prepend
my_array=("a" "${my_array[@]}")

Load an Array from a Space-Separated List

Method 1 - use a 'for' iterator:

local space_separated_list="a b c d"
    
declare -a my_array

for i in ${space_separated_list}; do
    my_array+=(${i})
done

echo "length: ${#my_array[@]}"
echo ${my_array[0]}
echo ${my_array[1]}
echo ${my_array[2]}
echo ${my_array[3]}


Method 2 - use read:

(note that this can be used for other separators as well, by setting IFS)

local space_separated_list="a b c d"
    
read -r -a my_array <<< "${space_separated_list}"

echo "length: ${#my_array[@]}"
echo ${my_array[0]}
echo ${my_array[1]}
echo ${my_array[2]}
echo ${my_array[3]}

Load an Array from Command-Line Positional Arguments

declare -a my_array
my_array=("$@")

Read the Content of a File into An Indexed Array

Each line will be assigned to an element of the array:

readarray < ./somefile.txt -t my_array_var

It DOES NOT split the file by lines! Only the first line will be loaded into the array, where each space separated element will be assigned to an array element:

declare -a VALUES
read -r -a VALUES < ./file.txt

Assign Words Read from stdin to an Indexed Array

Assing words read from stdin to an array:

read -a

Indexed Array Access

Dereference the Array Variable

Dereferencing the array variable returns the value of the first element of the array. For

declare -a VALUES
VALUES=("A" "B" "C")
echo ${VALUES}

produces "A".

Indexed Arrays Access by Index

Reference Individual Indexed Array Elements

Element arrays can be referenced with ${var-name[index]} notation:

echo ${a[0]}

If the array variable was not previously declared, or if the specific element was not initialized, a reference attempt will return the blank string.

Get All Indexed Array Elements

If "@" or * is used as subscript, the following expressions expand to all members of the array:

${some_array[*]}
"${some_array[@]}"

The difference between * and "@" relative to any array elements, in general, is described here:

* and @ Relative to Array Elements

Reference a Sub-Array

A contiguous sub-array can be referenced as follows:

${var_name[@]:from}

returns all elements of the array starting with the index 'from'.

${var_name[@]:from:to}

returns all elements of the array starting with the index 'from' up to index 'to'.

Indirect Variable Access for Indexed Arrays

It is possible to use variable indirection with indexed array variables. Knowing the name of the array variable, the values corresponding to a certain index can be accessed as follows:

declare -a VALUES
VALUES=("A" "B" "C")

The value of the first element of the array is displayed as follows:

name="VALUES[0]"
echo ${!name}

The value of the second element of the array is displayed as follows:

name="VALUES[1]"
echo ${!name}

Passing the Name of an Indexed Array as Argument to a Function

Does not work, the values can be accessed inside the function, but the changes do not surface outside the function.

It is possible to declare an indexed array outside a function, then pass the name of the array to the function, which then can refer the underlying array and use it: access it and mutate it. This is an effective way of "pass by reference".

Example:

declare -a SOME_ARRAY

function array_manipulator() {
  local array_name=$1
  array=${!array_name}
  array+=(a)
}

array_manipulator SOME_ARRAY

echo ${SOME_ARRAY[0]}} # unfortunately, it does not displays "a"

Length of an Indexed Array

${#var_name[@]}
${#var_name[*]}

If the array is sparse, this does not return the index of last element plus one, but just the number of non-empty elements.

Iterate over an Indexed Array

Using the Indexed Array Elements

for i in "${array_var[@]}"; do
   echo "${i}"
done

Iterate Over the Argument List

args=("$@")
for arg in ${args[@]}; do
  echo "${arg}"
done

or

for arg in "$@"; do
   echo "${arg}"
done

Using the Index

i=0
while [[ ${i} < ${#array_var[@]} ]]; do
  echo ${array_var[${i}]}
  ((i++))
done

Note that this does not work when the array is sparse, because ${#array_var[@]} only returns the number of non-empty elements, and not the array length.

Clear an Indexed Array

Note that re-declaring an indexed array does not clear it. To be cleared, the array needs to be unset and redeclared.

declare -a A
...
unset A
declare -a A

Associative Arrays

Associative Array Overview

Associative arrays can be used to implement sets and maps in bash. The most common usage as set is to insert elements whose subscript is identical with the value.

Associative Array Declaration


Associative arrays must be explicitly declare before initialization, otherwise unexpected behavior may occur. For example, if multiple assignments are declared, only the last one is reflected in the state of the array. This is different from the behavior of the indexed arrays, which can be used without explicit declaration.

declare -A some_assoc_array

This command does not create the associative array immediately, it just sets an attribute on the name some_assoc_array which allows assignments to the name as an associative array. The array does not exist until the first assignment. The fact that a specific array was declared can be tested by executing declare -A and grepping the result, as shown below.

All previously declared (and not yet unset) associative arrays can be listed with:

declare -A

An associative array variable can be "undeclared" with the unset builtin.

unset some_assoc_array

Associative Array Declaration Test

This sequence tests whether a variable was declared as an associative array. The test does not tell whether the associative array was actually allocated.

declare -A | grep -q "declare -A some_assoc_array" && \
 echo "'some_assoc_array' was declared as associative array" || \
 echo "'some_assoc_array' was not declared as associative array"

Associative Array Allocation Test

This sequence tests whether an associative array was actually allocated (can be empty) and associated with the variable.

[[ -v some_assoc_array[@] ]] && \
  echo "there is an array associated with 'some_assoc_array'" || \
  echo "there is NO array associated with 'some_assoc_array'"

The above detects allocation in both following cases:

some_assoc_array["a"]=
some_assoc_array["a"]="b"

Also see:

bash test -v

Associative Array Assignment

Assign Individual Associative Array Elements

Individual elements are initialized by using assoc_array_var_name[key]=value notation:

a["color"]="blue"

The associative array variable must be declared with declare -A in advance, otherwise only the last assignment is reflected in the state of the array.

Entire Associative Array

The ([key_1]=value_1 [key_2]=value_2 ...) syntax can be used to initialize an entire array. The values must be separated by space:

a=(["color]="red" ["size"]="small")

Associative Array Key Existence

The existence of a specific key in an associative array (whether it has a value associated with it or not) can be tested with:

declare -A A

[[ ${A["some-key"]+nonemptystring} ]] && echo "'some-key' exists" || echo "'some-key' does not exist"

The key is identified to exist even if it has an empty value. In the following cases, the key is identified as existent:

A["some-key"]=
A["some-key"]=""
A["some-key"]="something"

This should work for both associative arrays and indexed arrays.

More general:

declare -A A
key="somekey"
A[${key}]="somevalue"

[[ ${A[${key}]+nonemptystring} ]] && echo "'${key}' exists" || echo "'${key}' does not exist"

Associative Array Access

Reference Individual Associative Array Elements

Element arrays can be referenced with ${assoc_array_var_name[key]} notation:

echo ${some_assoc_array["color"]}

If the array variable was not previously declared, or if the specific element was not initialized, a reference attempt will return the blank string.

Get All Associative Array Keys

All keys are returned with:

${!some_assoc_array[*]}

or

"${!some_assoc_array[@]}"

The difference between !some_assoc_array[*] and "${!some_assoc_array[@]}" shows in the case the key contain spaces. In the former situation, the keys that contain spaces are tokenized into individual non-space-containing strings. In the later situation, the keys that contain spaces are handled as individual strings that contain spaces.

The difference between * and @ relative to any array elements, in general, is described here:

* and @ Relative to Array Elements

Get All Associative Array Values

If "@" or * is used as subscript, the following expressions expand to all values maintained in the array:

${array_var_name[@]}
${array_var_name[*]}

The difference between * and "@" relative to any array elements, in general, is described here:

* and @ Relative to Array Elements

Iterate over Key/Values of an Associative Array

for k in "${!array_var_name[@]}"; do
    echo "key: ${k}"
    echo "value: ${array_var_name[${k}]}"
done

Enclose ${!array_var_name[@]} in double quotes.

Indirect Variable Access for Associative Arrays

It is possible to use variable indirection with associative array variables. Knowing the name of the associative array variable, the values corresponding to a certain key can be accessed as follows:

declare -A VALUES
VALUES["color"]="red"

The value associated with the "color" key is displayed as follows:

name="VALUES[\"color\"]"
echo ${!name}

The value associated with the "color" key can be modified as follows:

name="VALUES[\"color\"]"
eval "${name}=blue"

This will produce:

echo ${VALUES["color"]}
blue

Also see:

eval

Size

${#var_name[@]}
${#var_name[*]}

Clear the Array

To clear an individual element of the array, use:

unset array_name[subscript]
declare -a a
a["color"]="blue"
unset a["color"]

To clear the entire array, but not discard the variable, iterate over the array and unset each element:

for k in "${!array_var_name[@]}"; do
   unset array_var_name[${k}]
done

Enclose ${!array_var_name[@]} in double quotes.

Associative Array Patterns