Differences between revisions 45 and 46
Revision 45 as of 2012-11-30 17:13:10
Size: 18502
Editor: ormaaj
Comment: read -d '' finally fixed.
Revision 46 as of 2012-12-01 15:17:42
Size: 18500
Editor: 148
Comment: fix toc
Deletions are marked like this. Additions are marked like this.
Line 3: Line 3:

How can I use array variables?

This answer assumes you have a basic understanding of what arrays are. If you're new to this kind of programming, you may wish to start with the guide's explanation. This page is more thorough. See links at the bottom for more resources.

1. Intro

One-dimensional integer-indexed arrays are implemented by Bash, Zsh, and most KornShell varieties including AT&T ksh88 or later, mksh, and pdksh. Arrays are not specified by POSIX and not available in legacy or minimalist shells such as BourneShell and Dash. The POSIX-compatible shells that do feature arrays mostly agree on their basic principles, but there are some significant differences in the details. Advanced users of multiple shells should be sure to research the specifics. Ksh93, Zsh, and Bash 4.0 additionally have Associative Arrays. This article focuses on indexed arrays as they are the most common and useful type.

Here is a typical usage pattern featuring an array named host:

# Bash

# Assign the values "mickey", "minnie", and "goofy" to sequential indexes starting with zero.
host=(mickey minnie goofy)

# Iterate over the indexes of "host".
for idx in "${!host[@]}"; do
    printf 'Host number %d is %s' "$idx" "${host[idx]}"
done

"${!host[@]}" expands to the indices of of the host array, each as a separate argument. (We'll go into more detail on syntax below.)

Indexed arrays are sparse, and elements may be inserted and deleted out of sequence.

# Bash/ksh

# Simple assignment syntax.
arr[0]=0
arr[2]=2
arr[1]=1
arr[42]='what was the question?'

# Unset the seceond element of "arr"
unset -v 'arr[2]'

# Concatenate the values, to a single argument separated by spaces, and echo the result.
echo "${arr[*]}"
# outputs: "0 1 what was the question?"

It is good practice to write your code in such a way that it can handle sparse arrays, even if you think you can guarantee that there will never be any "holes". Only treat arrays as "lists" if you're certain, and the savings in complexity is significant enough for it to be justified.

2. Loading values into an array

Assigning one element at a time is simple, and portable:

# Bash/ksh
arr[0]=0
arr[42]='the answer'

It's possible to assign multiple values to an array at once, but the syntax differs across shells. Bash supports only the arrName=(args...) syntax. ksh88 supports only the set -A arrName -- args... syntax. ksh93, mksh, and zsh support both. There are subtle differences in both methods between all of these shells if you look closely.

# Bash, ksh93, mksh, zsh
array=(zero one two three four)

# ksh88/93, mksh, zsh
set -A array -- zero one two three four

When initializing in this way, the first index will be 0 unless a different index is specified.

With compound assignment, the space between the parentheses is evaluated in the same way as the arguments to a command, including pathname expansion and WordSplitting. Any type of expansion or substitution may be used. All the usual quoting rules apply within.

# Bash/ksh93
oggs=(*.ogg)

With ksh88-style assignment using set, the arguments are just ordinary arguments to a command.

# Korn
set -A oggs -- *.ogg

# Bash (brace expansion requires 3.0 or higher)
homeDirs=(~{,root}) # brace expansion occurs in a different order in ksh, so this is bash-only.
letters=({a..z})    # Not all shells with sequence-expansion can use letters.

# Korn
set -A args -- "$@"

2.1. Loading lines from a file or stream

In bash 4, the mapfile command (also known as readarray) accomplishes this:

# Bash 4
mapfile -t lines <myfile

# or
mapfile -t lines < <(some command)

See ProcessSubstitution and FAQ #24 for more details on the <(...) syntax.

mapfile handles blank lines by inserting them as empty array elements, and also missing final newlines from the input stream. These can be problematic when reading data in other ways (see the next section). mapfile does have one serious drawback: it can only handle newlines as line terminators. Not all options supported by read are handled by mapfile, and visa-versa. mapfile can't, for example, handle NUL-delimited files from find -print0`. When mapfile isn't available, we have to work very hard to try to duplicate it. There are a great number of ways to almost get it right, but fail in subtle ways.

These examples will duplicate most of mapfile's basic functionality:

# Bash, Ksh93, mksh
while IFS= read -r; do
    lines+=("$REPLY")
done <file
[[ $REPLY ]] && lines+=("$REPLY")

The += operator, when used together with parentheses, appends the element to one greater than the current highest numbered index in the array.

# Korn
# Ksh88 doesn't support pre/post increment/decrement. mksh and others do.
i=0
while IFS= read -r; do
    lines[i+=1,$i]=$REPLY
done <file
[[ $REPLY ]] && lines[i]=$REPLY

The square brackets create a math context. The result of the expression is the index used for assignment.

2.1.1. Handling newlines (or lack thereof) at the end of a file

read returns false when it reads the last line of a file. This presents a problem: if the file contains a trailing newline, then read will be false when reading/assigning that final line, otherwise, it will be false when reading/assigning the last line of data. Without a special check for these cases, no matter what logic is used, you will always end up either with an extra blank element in the resulting array, or a missing final element.

To be clear - most text files should contain a newline as the last character in the file. Newlines are added to the ends of files by most text editors, and also by Here documents and Here strings. Most of the time, this is only an issue when reading output from pipes or process substitutions, or from "broken" text files created with broken or misconfigured tools. Let's look at some examples.

This approach reads the elements one by one, using a loop.

# Doesn't work correctly!
unset -v arr i
while IFS= read -r 'arr[i++]'; do
    :
done < <(printf '%s\n' {a..d})

Unfortunately, if the file or input stream contains a trailing newline, a blank element is added at the end of the array, because the read -r arr[i++] is executed one extra time after the last line containing text before returning false.

# Still doesn't work correctly!
unset -v arr i
while read -r; do
    arr[i++]=$REPLY
done < <(printf %s {a..c}$'\n' d)

The square brackets create a math context. Inside them, i++ works as a C programmer would expect (in all but ksh88).

This approach fails in the reverse case - it correctly handles blank lines and inputs terminated with a newline, but fails to record the last line of input. If the file or stream is missing its final newline. So we need to handle that case specially:

# Bash, ksh93, mksh
unset -v arr i
while IFS= read -r; do
    arr[i++]=$REPLY
done <file
[[ $REPLY ]] && arr[i++]=$REPLY # Append unterminated data line, if there was one.

This is very close to the "final solution" we gave earlier -- handling both blank lines inside the file, and an unterminated final line. The null IFS is used to prevent read from stripping possible whitespace from the beginning and end of lines, in the event you wish to preserve them.

Another workaround is to remove the empty element after the loop:

# Bash
unset -v arr i
while IFS= read -r 'arr[i++]'; do
    :
done <file

# Remove trailing empty element, if any.
[[ ${arr[i-1]} ]] || unset -v 'arr[--i]'

Whether you prefer to read too many and then have to remove one, or read too few and then have to add one, is a personal choice.

NOTE: it is necessary to quote the 'arr[i++]' passed to read, so that the square brackets aren't interpreted as globs. This is also true for other non-keyword builtins that take a subscripted variable name, such as let and unset.

2.1.2. Other methods

Sometimes stripping blank lines actually is desirable, or you may know that the input will always be newline delimited, such as input generated internally by your script. It is possible in some shells to use the -d flag to set read's line delimiter to null, then abuse the -a or -A (depending on the shell) flag normally used for reading the fields of a line into an array for reading lines. Effectively, the entire input is treated as a single line, and the fields are newline-delimited.

# Bash 4
    IFS=$'\n' read -rd '' -a lines <file

# mksh,  zsh
    IFS=$'\n' read -rd '' -A lines <file

Unfortunately, the above doesn't work in ksh93, even though its read does have the -d delimiter flag. Of course, the above examples do not preserve blank lines, but they are a quick easy mapfile replacement that also works in a few non-bash shells. Fixed as of ksh 93v- alpha 2012-10-12

12-10-09 +read -d '' now reads up to a NUL byte.

2.1.3. Don't read lines with for!

NEVER READ LINES USING for..in LOOPS! Relying on IFS WordSplitting causes issues if you have repeated whitespace delimiters, because they will be consolidated. It is not possible to preserve blank lines by having them stored as empty array elements this way. Even worse, special globbing chracters will be expanded without going to lengths to disable and then re-enable it. Just never use this approach - it is problematic, the workarounds are all ugly, and not all problems are solvable.

Because this is such an incredibly common mistake, below illustrates close to the best possible version of this hack, and how much harder it is than just doing it correctly -- and it still can't preserve consecutive newlines! It only gets worse from here. See DontReadLinesWithFor for details.

# Bash
# WARNING: Don't do this!

evilReadLines() {
    [[ -e $2 ]] || return

    # Try hard to preserve the previous glob and trap states.
    # But if the caller sets ERR or DEBUG, we're still in trouble!
    if [[ $- != *f* ]]; then
        set -f
        local oReturn=$(trap -p RETURN)
        trap 'set +f; trap "${oReturn:--}" RETURN' RETURN
    fi

    local line idx IFS=$'\n'
    for line in ${1:+$(<"$2")}; do
        printf -v "${1}[idx++]" %s "$line"
    done

    # This is an equally bad alternative to the above for loop, albeit slightly faster:
    # IFS=$'\n' declare -a ${1:+"$1"'=( $(<"$2") )'} 2>/dev/null
}
declare -ft evilReadLines # Inherit traps from the caller.

# Pass in an array name and file name
evilReadLines myArray myFile

2.2. Reading NUL-delimited streams

If you are trying to deal with records that might have embedded newlines, you will be using an alternative delimiter such as the NUL character ( \0 ) to separate the records. In that case, you'll need to use the -d argument to read as well:

# Bash
unset -v arr i
while IFS= read -rd '' 'arr[i++]'; do
    :
done < <(find . -name '*.ugly' -print0)

# or
while read -rd ''; do
    arr[i++]=$REPLY
done < <(find . -name '*.ugly' -print0)

# or (bash 3.1 and up)
while read -rd ''; do
    arr+=("$REPLY")
done < <(find . -name '*.ugly' -print0)

read -d '' tells Bash to keep reading until a NUL byte instead of until a newline. This isn't certain to work in all shells with a -d feature.

2.3. Appending to an existing array

As previously mentioned, arrays are sparse - that is, numerically adjacent indexes are not guaranteed to be occupied by a value. This confuses what it means to "append" to an existing array. There are several approaches.

If you've been keeping track of the highest-numbered index with a variable (for example, as a side-effect of populating an array in a loop), and can guarantee it's correct, you can just use it and continue to ensure it remains in-sync.

# Bash/ksh93
arr[++i]="new item"

If you don't want to keep an index variable, but happen to know that your array is not sparse, then you can use the number of elements to calculate the offset (not recommended):

# Bash/ksh
# This will FAIL if the array has holes (is sparse).
arr[${#arr[@]}]="new item"

If you don't know whether your array is sparse or not, but don't mind re-indexing the entire array (very inefficient), then you can use:

# Bash
arr=("${arr[@]}" "new item")

# Ksh
set -A arr -- "${arr[@]}" "new item"

If you're in bash 3.1 or higher, then you can use the += operator:

# Bash 3.1, ksh93, mksh, zsh
arr+=(item 'another item')

NOTE: the parentheses are required, just as when assigning to an array. Otherwise you will end up appending to ${arr[0]} which $arr is a synonym for. If your shell supports this type of appending, it is the preferred method.

For examples of using arrays to hold complex shell commands, see FAQ #50 and FAQ #40.

3. Retrieving values from an array

${#arr[@]} or ${#arr[*]} expand to the number of elements in an array:

# Bash
shopt -s nullglob
oggs=(*.ogg)
echo "There are ${#oggs[@]} Ogg files."

Single elements are retrieved by index:

echo "${foo[0]} - ${bar[j+1]}"

The square brackets are a math context. Within an arithmetic context, variables, including arrays, can be referenced by name. For example, in the expansion:

${arr[x[3+arr[2]]]}

arr's index will be the value from the array x whose index is 3 plus the value of arr[2].

Using array elements en masse is one of the key features of shell arrays. In exactly the same way that "$@" is expanded for positional parameters, "${arr[@]}" is expanded to a list of words, one array element per word. For example,

# Korn/Bash
for x in "${arr[@]}"; do
  echo "next element is '$x'"
done

This works even if the elements contain whitespace. You always end up with the same number of words as you have array elements.

If one simply wants to dump the full array, one element per line, this is the simplest approach:

# Bash/ksh
printf "%s\n" "${arr[@]}"

For slightly more complex array-dumping, "${arr[*]}" will cause the elements to be concatenated together, with the first character of IFS (or a space if IFS isn't set) between them. As it happens, "$*" is expanded the same way for positional parameters.

# Bash
arr=(x y z)
IFS=/; echo "${arr[*]}"; unset IFS
# prints x/y/z

Unfortunately, you can't put multiple characters in between array elements using that syntax. You would have to do something like this instead:

# Bash/ksh
arr=(x y z)
tmp=$(printf "%s<=>" "${arr[@]}")
echo "${tmp%<=>}"    # Remove the extra <=> from the end.
# prints x<=>y<=>z

Or using array slicing, described in the next section.

# Bash/ksh
typeset -a a=([0]=x [5]=y [10]=z)
printf '%s<=>' "${a[@]::${#a[@]}-1}"
printf '%s\n' "${a[@]:(-1)}"

This also shows how sparse arrays can be assigned multiple elements at once. Note using the arr=([key]=value ...) notation differs between shells. In ksh93, this syntax gives you an associative array by default unless you specify otherwise, and using it requires that every value be explicitly given an index, unlike bash, where omitted indexes begin at the previous index. This example was written in a way that's compatible between the two.

BASH 3.0 added the ability to retrieve the list of index values in an array:

# Bash 3.0 or higher
arr=(0 1 2 3) arr[42]='what was the question?'
unset 'arr[2]'
echo "${!arr[@]}"
# prints 0 1 3 42

Retrieving the indices is extremely important for certain kinds of tasks, such as maintaining parallel arrays with the same indices (a cheap way to mimic having an array of structs in a language with no struct):

# Bash 3.0 or higher
unset file title artist i
for f in ./*.mp3; do
  file[i]=$f
  title[i]=$(mp3info -p %t "$f")
  artist[i++]=$(mp3info -p %a "$f")
done

# Later, iterate over every song.
# This works even if the arrays are sparse, just so long as they all have
# the SAME holes.
for i in "${!file[@]}"; do
  echo "${file[i]} is ${title[i]} by ${artist[i]}"
done

3.1. Retrieving with modifications

Bash's Parameter Expansions may be performed on array elements en masse:

# Bash
arr=(abc def ghi jkl)
echo "${arr[@]#?}"          # prints bc ef hi kl
echo "${arr[@]/[aeiou]/}"   # prints bc df gh jkl

Parameter Expansion can also be used to extract elements from an array. Some people call this slicing:

# Bash
echo "${arr[@]:1:3}"        # three elements starting at #1 (second element)
echo "${arr[@]:(-2)}"       # last two elements
echo "${@:(-1)}"            # last positional parameter
echo "${@:(-2):1}"          # second-to-last positional parameter

4. Using @ as a pseudo-array

As we see above, the @ array (the array of positional parameters) can be used almost like a regularly named array. This is the only array available for use in POSIX or Bourne shells. It has certain limitations: you cannot individually set or unset single elements, and it cannot be sparse. Nevertheless, it still makes certain POSIX shell tasks possible that would otherwise require external tools:

# POSIX
set -- *.mp3
if [ -e "$1" ]; then
  echo "there are $# MP3 files"
else
  echo "there are 0 MP3 files"
fi

# POSIX
...
# Add an option to our dynamically generated list of options
set -- "$@" -f "$somefile"
...
foocommand "$@"

(Compare to FAQ #50's dynamically generated commands using named arrays.)

See Also


CategoryShell

BashFAQ/005 (last edited 2023-03-25 22:39:06 by emanuele6)