eolas/zk/Split_into_array_in_Bash.md

---
tags:
  - shell
---

# Splitting input into an array

## readarray

`readarray` makes it really easy to split input into an array based on new
lines. Say we have this file as input:

```
123
456
789
```

Then we can split like so:

```bash
readarray -t name_for_array < ./input.text

# Print all elements
echo "${name_for_array[@]}"

# Print element by index

echo "${name_for_array[1]}"
456
```

If we want to read a string directly:

```bash
readarray -t new_name_for_array <<< "here
is
some
text"

echo "${new_name_for_array[1]}"
is
```

> The _-t_ flag removes the trailing newline

See more: https://linuxhint.com/split-string-array-bash/

## read

For delimiters other than a space we have to use `read`, combined with `IFS` the
**Internal Field Separator**.

For example, to split by comma:

```plaintext
# comma-input.txt
something,something else,something more
```

```bash
IFS=',' read -a arr < ./comma_inputs.txt
```

> We use the `-a` flag to signal that we want each item read returned as an
> element in an array

## mapfile

We can use `mapfile` to read the lines of a input file in set segments.

For the following file:

```
line one
line two
line three
line four
```

We could generate the following output:

```
-----SNIP-----
line one
line two
-----SNIP-----
line three
line four
```

With:

```bash
while mapfile -t -n 2 ary && ((${#ary[@]})); do
  printf '%s\n' "${ary[@]}"
  printf -- '--- SNIP ---\n'
 done < $testInput
```

So, while there are input lines, map each line into an array and return it in
groups of 2.