Data Structures
Array
Bash has 1D indexed array (the common array we normally see), and associative array. Any variable can be used as an array. To declare an array, use declare.
Bash Array
An indexed array is “common”. However, it doesn’t guarantee that items are stored contiguously.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
#!/bin/bash
declare -a array1 array2 # one can declare multiple variables
array1=("apple" "peach")
echo ${array1[1]} # see apple,
echo ${array1[@]} # see apple,peach
array2=("pear")
# $array2 is the first element in array2 only, similar to c.
# array1+=$array2 #applepear peach
# This is also applepear peach
# array1+="${array2[@]}"
array1+=("${array2[@]}") #see apple, peach, pear ?
echo ${array1[@]} # see apple,peach
array1+=("pecan")
echo ${array1[@]} #see apple, peach, pear, pecan
Associative Array
A bash associative array is similar to a python dictionary. E.g., below is an excerpt of a script that downloads ROS bags
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
declare -A bags_link_lookup #?
# This is very dictionary-like
bags_link_lookup["data/rgbd_dataset_freiburg1_xyz.bag"]="https://cvg.cit.tum.de/rgbd/dataset/freiburg1/rgbd_dataset_freiburg1_xyz.bag"
bags_link_lookup["data/rgbd_dataset_freiburg1_rpy.bag"]="https://cvg.cit.tum.de/rgbd/dataset/freiburg1/rgbd_dataset_freiburg1_rpy.bag"
echo "Download bag files if they are missing ..."
for key in "${!bags_link_lookup[@]}"; do
if [[ -f "${key}" ]]; then
echo "No need to download dataset ${key}..."
else
echo "Downloading dataset ${key}"
wget -P data "${bags_link_lookup[$key]}"
fi
done
echo "Done Downloading"
Commands
Find
- find regular files and count their numbers:
find . -type f | wc -lfind . -type flists all regular files recursively from the current directory (.)wc -lcounts the number of lines
Tree
1
- `tree -L 2`: limits the search depth to 2
Grep
grep -B 5 "Something": grep the next 5 lines of each instancegrep -Rinvsgrep -rin:-rdoesn’t allow softlink.
Compound Command?
A “compound command” in Bash is any control‐flow construct that groups multiple simple commands into a single logical unit. Examples of compound commands include:
- Loops:
for …; do …; done - Conditionals:
if …; then …; fi - Grouped commands:
{ …; }or( … ) - Case statements:
case …; esac
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
LIST="$(mktemp)"
# Works
for f in "$DIR"/*.mp4; do
echo "Found: $f" >&2
[[ -e "$f" ]] || continue
printf "file '%s'\n" "$f"
done > "$LIST"
# DOESN'T WORK
for f in "$DIR"/*.mp4; do
echo $f
# Skip if no matches
[[ -e "$f" ]] || continue
# FFmpeg concat demuxer wants paths in single quotes
printf "file '%s'\n" "$f"
done > "$LIST"
This is because:
1
2
3
for f in "$DIR"/*.mp4; do
...
done > "$LIST"
the> "$LIST" redirection is attached to the entire for …; do …; done block. All standard‐output (stdout) from anything inside that loop — every echo, printf, or other command that writes to stdout—gets sent into "$LIST"
Every write to the stdout will go to file $LIST. You can either redirect it to stderr >&2, or check out the file: cat $LIST
tee
tee writes to standard input and one more file.
Bind
- Create a custom keyboard shortcut that triggers a command in shell: (in this case,
navi)
1
bind -x '"\C-f": navi'
Command
command -v lcov: command -v checks if this is a command is an executable;
1
if command -v lcov > /dev/null # makes it go to null
Variables
IFS
IFS: IFS is a special built-in shell variable that controls how Bash splits words. Many commands rely on it implicitly, but the most common place you’ll see it used intentionally is with the read builtin. By default, IFS contains:
1
space, tab, newline
That means when you run:
1
read -ra arr <<< "a b c"
Bash splits the input on whitespace and populates the array accordingly.
Why IFS Matters
The read command is hardcoded to use the variable named exactly IFS for field splitting. You cannot rename it. If you want to change how input is split, you must modify IFS. For example, suppose you have a pipe-delimited string:
1
OVERLAPS="foo|bar|baz"
You can split it like this:
1
IFS='|' read -ra OVERLAP_ARRAY <<< "$OVERLAPS"