{"id":3508,"date":"2026-10-01T18:15:01","date_gmt":"2026-10-01T18:15:01","guid":{"rendered":"https:\/\/www.ucartz.com\/blog\/bash-for-loop\/"},"modified":"2026-10-01T18:15:11","modified_gmt":"2026-10-01T18:15:11","slug":"bash-for-loop","status":"publish","type":"post","link":"https:\/\/www.ucartz.com\/blog\/bash-for-loop\/","title":{"rendered":"Bash for Loop: Syntax and Examples for Server Admins"},"content":{"rendered":"\n

Short answer:<\/strong> a Bash for<\/code> loop runs the same commands once for each item in a list: for item in one two three; do echo \"$item\"; done<\/code>. Use {1..10}<\/code> for a fixed range of numbers, for (( i = 1; i <= n; i++ ))<\/code> when you need a counter, and for f in *.log<\/code> to work through files. Every example below was run in Bash 5.2 on Debian 13, and the output shown is what we got.<\/p>\n\n\n\n

\"Terminal
Bash for loops run on a Debian 13 server: a word list, {1..5}, a C-style counter and a loop over files.<\/figcaption><\/figure>\n\n\n\n

The basic syntax<\/h2>\n\n\n\n

The list form is the one you will use most. Bash expands the words after in<\/code>, sets the variable to each word in turn, and runs everything between do<\/code> and done<\/code>:<\/p>\n\n\n\n

for name in alpha beta gamma; do\n  echo \"Hello, $name\"\ndone<\/code><\/pre>\n\n\n\n
Hello, alpha\nHello, beta\nHello, gamma<\/code><\/pre>\n\n\n\n
You can write the same loop on one line. The semicolons stand in for the line breaks:<\/p>\n\n\n\n
for i in 1 2 3; do echo \"Line $i\"; done<\/code><\/pre>\n\n\n\n
Two details from the Bash manual<\/a> are worth knowing. If you leave out in ...<\/code> completely, the loop runs over the script’s arguments (\"$@\"<\/code>). And if the list expands to nothing, the body never runs and the loop returns status 0.<\/p>\n\n\n\n
Loop over a range of numbers<\/h2>\n\n\n\n
Brace expansion builds the list for you. You can add a step, pad with zeros or count through letters:<\/p>\n\n\n\n
for i in {1..5}; do echo \"Number $i\"; done      # 1 to 5\nfor i in {0..20..5}; do echo -n \"$i \"; done     # 0 5 10 15 20\nfor i in {01..10}; do echo -n \"$i \"; done       # 01 02 ... 10\nfor c in {a..e}; do echo -n \"$c \"; done         # a b c d e<\/code><\/pre>\n\n\n\n
There is one trap here. Brace expansion happens before<\/em> variable expansion, so a variable inside the braces does not work:<\/p>\n\n\n\n
n=5\nfor i in {1..$n}; do echo \"$i\"; done<\/code><\/pre>\n\n\n\n
{1..5}<\/code><\/pre>\n\n\n\n
The loop ran once with the literal text {1..5}<\/code>. When the end of the range comes from a variable, use seq<\/code> or the C-style loop below:<\/p>\n\n\n\n
for i in $(seq 1 \"$n\"); do echo -n \"$i \"; done   # 1 2 3 4 5<\/code><\/pre>\n\n\n\n
The C-style for loop<\/h2>\n\n\n\n
Bash also has an arithmetic form with a start, a test and a step, like C or JavaScript. It reads variables directly, so it is the cleanest choice for counters:<\/p>\n\n\n\n
n=5\nfor (( i = 1; i <= n; i++ )); do\n  echo -n \"$i \"\ndone\n# 1 2 3 4 5\n\nfor (( i = 10; i > 0; i -= 3 )); do echo -n \"$i \"; done\n# 10 7 4 1<\/code><\/pre>\n\n\n\n
Loop over files<\/h2>\n\n\n\n
Put a wildcard pattern in the list and Bash replaces it with the matching file names, sorted. Always quote the variable, because file names can contain spaces:<\/p>\n\n\n\n
for f in *.log; do\n  echo \"Found: $f\"\ndone<\/code><\/pre>\n\n\n\n
Found: access.log\nFound: error.log\nFound: my site.log<\/code><\/pre>\n\n\n\n
Do not loop over the output of ls<\/code>. In the same folder, for f in $(ls *.log)<\/code> printed my<\/code> and site.log<\/code> as two separate files, because the output is split on spaces.<\/p>\n\n\n\n
Also watch what happens when nothing matches. Without extra options, the pattern stays as plain text, so this loop runs once with a file name that does not exist:<\/p>\n\n\n\n
for f in *.gz; do echo \"Found: $f\"; done\n# Found: *.gz<\/code><\/pre>\n\n\n\n
Turn on nullglob<\/code> and a pattern with no matches expands to nothing, so the loop simply does not run:<\/p>\n\n\n\n
shopt -s nullglob\nfor f in *.gz; do echo \"Found: $f\"; done\n# (no output)<\/code><\/pre>\n\n\n\n
Loop over an array<\/h2>\n\n\n\n
Use \"${array[@]}\"<\/code> with the quotes, so an item with a space stays one item. To get the index as well, loop over \"${!array[@]}\"<\/code>:<\/p>\n\n\n\n
servers=(\"web1\" \"web2\" \"db 1\")\n\nfor s in \"${servers[@]}\"; do\n  echo \"Server: $s\"\ndone\n\nfor i in \"${!servers[@]}\"; do\n  echo \"$i: ${servers[$i]}\"\ndone<\/code><\/pre>\n\n\n\n
Server: web1\nServer: web2\nServer: db 1\n0: web1\n1: web2\n2: db 1<\/code><\/pre>\n\n\n\n
Reading a file line by line: use while, not for<\/h2>\n\n\n\n
A common mistake is for line in $(cat file)<\/code>. It splits on every space, not on line breaks. With a file containing example.com<\/code>, example.org<\/code> and my site.test<\/code>, it gave four items instead of three. A while read<\/code> loop keeps each line whole:<\/p>\n\n\n\n
while IFS= read -r domain; do\n  echo \"[$domain]\"\ndone < domains.txt<\/code><\/pre>\n\n\n\n
[example.com]\n[example.org]\n[my site.test]<\/code><\/pre>\n\n\n\n
Skip or stop: continue and break<\/h2>\n\n\n\n
continue<\/code> jumps to the next item and break<\/code> leaves the loop. This prints odd numbers and stops after 7:<\/p>\n\n\n\n
for i in {1..10}; do\n  if (( i % 2 == 0 )); then continue; fi\n  if (( i > 7 )); then break; fi\n  echo -n \"$i \"\ndone\n# 1 3 5 7<\/code><\/pre>\n\n\n\n
For more on the if<\/code> tests used here, see Understanding Bash if…else statements<\/a>.<\/p>\n\n\n\n
Practical examples for servers<\/h2>\n\n\n\n
Check the HTTP status of several URLs<\/h3>\n\n\n\n
for url in https:\/\/example.com\/ https:\/\/example.com\/missing-page; do\n  code=$(curl -s -o \/dev\/null -w \"%{http_code}\" \"$url\")\n  echo \"$code  $url\"\ndone<\/code><\/pre>\n\n\n\n
200  https:\/\/example.com\/\n404  https:\/\/example.com\/missing-page<\/code><\/pre>\n\n\n\n
Look up the IP address of several domains<\/h3>\n\n\n\n
for d in example.com example.org; do\n  echo \"$d: $(dig +short A \"$d\" | head -n1)\"\ndone<\/code><\/pre>\n\n\n\n
This prints each domain with its first IPv4 address. See DNS lookup commands<\/a> for more dig<\/code> options.<\/p>\n\n\n\n
Compress log files and rename extensions<\/h3>\n\n\n\n
for f in *.log; do gzip -- \"$f\"; done\n\nfor f in *.jpeg; do mv -- \"$f\" \"${f%.jpeg}.jpg\"; done<\/code><\/pre>\n\n\n\n
The --<\/code> stops a file name that starts with a dash from being read as an option. ${f%.jpeg}<\/code> removes .jpeg<\/code> from the end of the name. In our test, c d.jpeg<\/code> became c d.jpg<\/code> without trouble because the variables are quoted.<\/p>\n\n\n\n
Run jobs in parallel and wait for all of them<\/h3>\n\n\n\n
for i in {1..3}; do\n  sleep 1 &\ndone\nwait\necho \"all done\"<\/code><\/pre>\n\n\n\n
The &<\/code> starts each command in the background and wait<\/code> pauses until they have all finished. Keep the number of parallel jobs small on a busy server.<\/p>\n\n\n\n
Quick reference<\/h2>\n\n\n\n
You want to loop over<\/th>Use<\/th>Watch out for<\/th><\/tr><\/thead>
A few fixed words<\/td>for x in a b c<\/code><\/td>Quote \"$x\"<\/code> when you use it<\/td><\/tr>
A fixed range<\/td>for i in {1..10}<\/code><\/td>No variables inside the braces<\/td><\/tr>
A range from a variable<\/td>for (( i = 1; i <= n; i++ ))<\/code><\/td>Or $(seq 1 \"$n\")<\/code><\/td><\/tr>
Files<\/td>for f in *.log<\/code><\/td>Never $(ls)<\/code>; consider shopt -s nullglob<\/code><\/td><\/tr>
An array<\/td>for s in \"${arr[@]}\"<\/code><\/td>Keep the quotes and the [@]<\/code><\/td><\/tr>
Lines of a file<\/td>while IFS= read -r line<\/code><\/td>for<\/code> splits on spaces<\/td><\/tr>
Script arguments<\/td>for arg<\/code> (no in<\/code>)<\/td>Same as for arg in \"$@\"<\/code><\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
Before you run a loop on a live server<\/h2>\n\n\n\n