Reading Command Output Into an Array

#!/bin/bash
data=$'apple\nbanana\ncherry'
mapfile -t fruits <<< "$data"
echo "Element count: ${#fruits[@]}"
echo "Second fruit: ${fruits[1]}"

#!/bin/bash
printf "line one\nline two\nline three\n" > lines.txt
readarray -t lines < lines.txt
for line in "${lines[@]}"; do
    echo "Read: $line"
done

#!/bin/bash
data=$'first\nsecond'
mapfile -t clean_lines <<< "$data"
mapfile raw_lines <<< "$data"
echo "Clean (with -t): [${clean_lines[0]}]"
echo "Raw (without -t): [${raw_lines[0]}]"

#!/bin/bash
read -a fields <<< "red green blue"
echo "Number of fields: ${#fields[@]}"
echo "Second field: ${fields[1]}"

1. Using arr=($(command)) and assuming each line becomes one element; that form actually word-splits on all whitespace (not just newlines), so a line with multiple words gets broken into multiple array elements.
2. Forgetting mapfile/readarray (they are the same command, readarray is an alias) are Bash 4.0+ features and won't work on very old Bash or on sh.
3. Not using -t with mapfile/readarray, which leaves the trailing newline character attached to the end of every captured element.

• mapfile -t arr <<< "$data" (or < file) reads input line by line directly into an array, one element per line.
• readarray is simply another name for the exact same builtin as mapfile -- they are fully interchangeable.
• The -t flag strips the trailing newline from each line before storing it as an element -- almost always what you want.
• read -a arr <<< "$line" splits a single line into an array by whitespace (using IFS), which is a different use case from reading multiple lines.