Shell Scripting Complete 5: Bash Loops with for, while, until, while read and IFS

By Ravindra Bagale, trainer · LinkedIn · Last updated: 25 October 2026

1. An everyday example: a teacher taking attendance

Why do we need this?

A teacher opens the register and goes name by name: read a name, call it, mark P or A, move to the next, stop at the end. If a student has left the school, she skips the name. If the bell rings she stops early. That is exactly a while-read loop with continue and break.

What is it?

Register = the input file. One name per line = one line per loop. Columns roll no, name, class = fields split by IFS. Skip a name = continue. Bell rings = break. End of register = read returns false.

How to do it (step by step)

2. for over a list

Why do we need this?

The simplest way to repeat the same command for a few known items: servers, users, environments.

What is it?

for name in item1 item2 item3; do ...; done. Each loop name holds the next item. The list can be words, "$@", a glob like *.csv, a brace range {1..5}, or $(command) (careful: splits on spaces).

How to do it (step by step)

1. Words.

for env in dev stage prod; do echo "deploying to $env"; done

Output:

deploying to dev
deploying to stage
deploying to prod

What it means: Three items, three runs.

2. Brace range.

for i in {1..3}; do echo "day $i"; done

Output:

day 1
day 2
day 3

What it means: {1..3} makes 1 2 3.

3. Step and zero pad.

echo {01..10..3}

Output:

01 04 07 10

What it means: Start, end, step.

4. Arguments.

set -- 'my file' other; for a in "$@"; do echo "[$a]"; done

Output:

[my file]
[other]

What it means: "$@" keeps spaces.

3. for over files with globs

Why do we need this?

Bulk work on files is the most common loop: rename, compress, convert, upload.

What is it?

for f in *.log handles names with spaces correctly (do not use $(ls)). Quote "$f" inside. If nothing matches, bash keeps the pattern itself (*.log); add shopt -s nullglob so the loop simply runs zero times. For sub-folders use shopt -s globstar and **/*.log, or find.

How to do it (step by step)

1. Compress logs.

for f in /tmp/demo/*.log; do gzip -k "$f" && echo "zipped $f"; done

Output:

zipped /tmp/demo/app.log
zipped /tmp/demo/web server.log

What it means: The name with a space works.

2. Nothing matches.

for f in /tmp/none/*.csv; do echo "got $f"; done

Output:

got /tmp/none/*.csv

What it means: The pattern itself came through.

3. nullglob fixes it.

shopt -s nullglob; for f in /tmp/none/*.csv; do echo "got $f"; done; echo done

Output:

done

What it means: Zero runs.

4. Change extensions.

for f in *.jpeg; do mv -- "$f" "${f%.jpeg}.jpg"; done; ls

Output:

a.jpg  b.jpg

What it means: -- protects names starting with -.

4. C-style for and seq

Why do we need this?

When you need a counter with a variable end (for example $n retries), brace ranges do not work.

What is it?

for (( i=1; i<=n; i++ )); do ...; done works like C or Java: start, condition, step. {1..$n} does NOT work because braces expand before variables. seq 1 $n also works and supports decimals and -w for zero padding.

How to do it (step by step)

1. Counter.

n=3; for (( i=1; i<=n; i++ )); do echo "try $i of $n"; done

Output:

try 1 of 3
try 2 of 3
try 3 of 3

What it means: Variables work here.

2. Braces fail.

n=3; echo {1..$n}

Output:

{1..3}

What it means: Not expanded.

3. seq.

for i in $(seq -w 8 10); do echo "file_$i"; done

Output:

file_08
file_09
file_10

What it means: -w pads to equal width.

4. Count down.

for (( i=3; i>0; i-- )); do echo $i; sleep 1; done; echo go

Output:

3
2
1
go

What it means: i-- goes down.

5. while and until

Why do we need this?

Sometimes you do not know how many times to loop: wait until a port opens, retry until a download works.

What is it?

while COMMAND; do ...; done repeats while the command returns 0. until COMMAND repeats while it fails (until it works). Always have a way out: a counter or a timeout, so the loop cannot run forever. while true; do ...; done is an endless loop you leave with break.

How to do it (step by step)

1. Wait for a port.

tries=0; until nc -z localhost 8080 || (( ++tries > 5 )); do echo "waiting ($tries)"; sleep 2; done

Output:

waiting (1)
waiting (2)

What it means: Stops when the port opens or after 5 tries.

2. While counter.

n=1; while (( n <= 3 )); do echo "round $n"; ((n++)); done

Output:

round 1
round 2
round 3

What it means: Condition checked each time.

3. Retry a download.

for i in 1 2 3; do curl -fsS -o /tmp/p.html https://example.com && break; echo "retry $i"; sleep 3; done; ls -l /tmp/p.html

Output:

-rw-r--r-- 1 asha asha 1256 Oct 25 /tmp/p.html

What it means: break leaves on success.

4. Watch a file grow.

while sleep 5; do wc -l < /var/log/syslog; done

Output:

3410
3415

What it means: Ctrl+C to stop.

6. break and continue

Why do we need this?

Real data has bad rows to skip and moments to stop early.

What is it?

continue skips the rest of this round and starts the next. break leaves the loop completely. In nested loops break 2 and continue 2 act on the outer loop.

How to do it (step by step)

1. Skip comments.

printf 'web1\n#old\nweb2\n' | while read -r h; do [[ $h == \#* ]] && continue; echo "ping $h"; done

Output:

ping web1
ping web2

What it means: Lines starting with # are skipped.

2. Stop at the first error.

for f in a.txt b.txt c.txt; do [[ -f $f ]] || { echo "missing $f, stopping"; break; }; echo "ok $f"; done

Output:

ok a.txt
missing b.txt, stopping

What it means: break leaves the loop.

3. Nested break 2.

for r in 1 2; do for c in a b; do [[ $c == b ]] && break 2; echo $r$c; done; done

Output:

1a

What it means: Both loops ended.

7. while IFS= read -r: read a file line by line

Why do we need this?

Host lists, user lists and exports are line-based files. This is the correct, safe way to process them. for line in $(cat file) splits on every space and is wrong.

What is it?

while IFS= read -r line; do ...; done < file. IFS= stops trimming of leading and trailing spaces. -r stops backslash processing. < file at the end feeds the loop. If the last line has no new line, add || [[ -n $line ]] so it is not lost. Use read -u 3 with 3< file when the body also reads from the keyboard or runs ssh.

How to do it (step by step)

1. Basic.

printf '  indented line\nC:\\path\n' | tee demo.txt; while IFS= read -r line; do echo "[$line]"; done < demo.txt

Output:

[  indented line]
[C:\path]

What it means: Spaces and backslash kept.

2. Wrong way.

for line in $(cat demo.txt); do echo "[$line]"; done

Output:

[indented]
[line]
[C:\path]

What it means: Split on spaces.

3. Last line without new line.

printf 'a\nb' | while IFS= read -r l || [[ -n $l ]]; do echo $l; done

Output:

a
b

What it means: b is not lost.

4. ssh eats stdin.

while read -r h; do ssh -n "$h" uptime; done < hosts.txt

Output:

web1: 10:00 up 3 days
web2: 10:00 up 9 days

What it means: -n stops ssh reading the list.

8. IFS: split fields (CSV, /etc/passwd)

Why do we need this?

Most data has columns separated by a comma, colon or tab. read can split them into named variables.

What is it?

IFS=, read -r a b c splits the line on commas into a, b and c. The last variable gets the rest of the line. Setting IFS before read changes it only for that read. Use IFS=: for /etc/passwd and IFS=$'\t' for tab files. For CSVs with quoted commas inside fields, use a real CSV tool (python, csvkit, mlr).

How to do it (step by step)

1. CSV.

printf 'id,name,city\n1,Asha,Pune\n2,Rohit,Satara\n' | tee people.csv; tail -n +2 people.csv | while IFS=, read -r id name city; do echo "$name lives in $city"; done

Output:

Asha lives in Pune
Rohit lives in Satara

What it means: tail -n +2 skips the header.

2. passwd.

while IFS=: read -r user _ uid _ _ home shell; do (( uid >= 1000 && uid < 65000 )) && echo "$user $home"; done < /etc/passwd

Output:

asha /home/asha
rohit /home/rohit

What it means: _ throws away columns.

3. Rest of the line.

IFS=: read -r a rest <<< 'one:two:three'; echo "$a | $rest"

Output:

one | two:three

What it means: The last variable takes the rest.

4. Tab separated.

printf 'sku\tqty\nA1\t5\n' | while IFS=$'\t' read -r s q; do echo "$s=$q"; done

Output:

sku=qty
A1=5

What it means: $'\t' is a tab.

9. The pipe subshell trap and process substitution

Why do we need this?

A very common surprise: you count inside a loop fed by a pipe and the count is 0 afterwards.

What is it?

In cmd | while read ...; done the loop runs in a subshell, so variables set inside are lost after done. Fix 1: feed with < file. Fix 2: process substitution done < <(cmd). Fix 3: shopt -s lastpipe in scripts. Process substitution <(cmd) makes a command look like a file, also useful for diff <(a) <(b).

How to do it (step by step)

1. The trap.

n=0; printf 'a\nb\nc\n' | while read -r x; do ((n++)); done; echo "count=$n"

Output:

count=0

What it means: The loop's n was in a subshell.

2. Fixed.

n=0; while read -r x; do ((n++)); done < <(printf 'a\nb\nc\n'); echo "count=$n"

Output:

count=3

What it means: The loop now runs in the main shell.

3. Compare two lists.

diff <(ls /etc | head -3) <(ls /etc | head -4)

Output:

3a4
> apt

What it means: Each command acts like a file.

10. Speed: parallel loops, wait and xargs -P

Why do we need this?

Checking 200 servers one by one is slow. Running several at once can cut the time a lot.

What is it?

Put & after a command inside the loop to run it in the background, then wait after the loop for all of them. Limit how many run together, for example with xargs -P 8 or by waiting every N jobs. Write each job's output to its own file so lines do not mix.

How to do it (step by step)

1. Background and wait.

for h in web1 web2 web3; do ( ping -c1 -W1 "$h" &>/dev/null && echo "$h up" || echo "$h down" ) & done; wait; echo all done

Output:

web2 up
web1 up
web3 down
all done

What it means: Order changes; all finish before 'all done'.

2. xargs -P.

printf '%s\n' web{1..6} | xargs -P 3 -I{} sh -c 'sleep 1; echo {} checked'

Output:

web1 checked
web2 checked
web3 checked
web4 checked
web5 checked
web6 checked

What it means: Three at a time, about 2 seconds total.

3. Time it.

time (for i in 1 2 3; do sleep 1 & done; wait)

Output:

real 0m1.01s

What it means: Three 1-second jobs in 1 second.

11. Real examples from many industries

Why do we need this?

The same Linux skill shows up in every kind of company. Seeing it in many places helps you remember it.

What is it?

Each row is a made-up team. It shows who uses this and how.

How to do it (step by step)

12. Common errors and how to fix them

Why do we need this?

Everyone hits these errors. Knowing the cause saves hours.

What is it?

Read the exact message first. Then match it to a row below.

How to do it (step by step)

13. Troubleshooting step by step

Why do we need this?

Make the loop show each value before acting on it.

What is it?

Go top to bottom. Stop at the first step that shows the problem.

How to do it (step by step)

1. Show items.

for f in *.log; do printf '[%s]\n' "$f"; done

2. Count lines read.

wc -l < file

3. Find CR characters.

grep -c $'\r' file

4. Strip CR.

sed -i 's/\r$//' file

5. Trace one round.

bash -x s.sh 2>&1 | head -40

6. Lint.

shellcheck s.sh

14. Lab 1: Users from a CSV

Why do we need this?

Read a CSV of new joiners with while IFS=, read -r, skip bad rows, and print the commands you would run.

What is it?

A full hands-on lab. Run every step on a test Linux machine (an EC2 instance, a VM, or WSL).

How to do it (step by step)

1. Folder and data.

mkdir -p ~/lab41 && cd ~/lab41 && printf 'user,team,shell\nmeera,dev,bash\n#skip,me,bash\n,ops,bash\nkunal,ops,zsh\n' | tee joiners.csv

2. Loop.

n=0; while IFS=, read -r u team sh; do [[ $u == \#* || -z $u ]] && continue; echo "useradd -m -s /bin/$sh -G $team $u"; ((n++)); done < <(tail -n +2 joiners.csv); echo "total $n"

Output:

useradd -m -s /bin/bash -G dev meera
useradd -m -s /bin/zsh -G ops kunal
total 2

3. Why total works.

echo 'process substitution kept the loop in this shell'

Output:

process substitution kept the loop in this shell

4. Check for Windows line ends.

grep -c $'\r' joiners.csv

Output:

0

5. Count rows.

tail -n +2 joiners.csv | wc -l

Output:

4

6. Lint idea.

echo 'run shellcheck on your script'

Output:

run shellcheck on your script

15. Lab 2: Retry and parallel check

Why do we need this?

Write a retry loop with until, then check many hosts in parallel with wait.

What is it?

A full hands-on lab. Run every step on a test Linux machine (an EC2 instance, a VM, or WSL).

How to do it (step by step)

1. Fake flaky command.

printf '#!/bin/bash\n(( RANDOM %% 3 == 0 ))\n' | tee flaky.sh && chmod +x flaky.sh

2. Retry.

t=0; until ./flaky.sh; do ((t++)); echo "failed $t"; (( t >= 10 )) && break; sleep 1; done; echo "finished after $t failures"

Output:

failed 1
failed 2
finished after 2 failures

3. Hosts file.

printf 'localhost\n127.0.0.1\nno.such.host\n' | tee hosts.txt

4. Parallel.

while IFS= read -r h; do ( ping -c1 -W1 "$h" &>/dev/null && echo "$h up" || echo "$h down" ) & done < hosts.txt; wait

Output:

localhost up
127.0.0.1 up
no.such.host down

5. Time it.

time (while IFS= read -r h; do ping -c1 -W1 "$h" &>/dev/null & done < hosts.txt; wait)

Output:

real 0m1.02s

6. xargs version.

xargs -P 3 -I{} ping -c1 -W1 {} < hosts.txt | grep -c 'bytes from'

Output:

2

16. Frequently asked questions

17. Interview questions and answers

Beginner level

Mid level level

Senior level

18. Related chapters

• Linux Ops Refresh for DevOps
• Bash Scripting for Automation
• Linux grep and find Commands with Examples
• How to Set Up Passwordless SSH Between Linux Servers

Revision, comparison and interview cards

Revise this lesson in a few minutes. Save the cards on your phone or download them as PNG files. The QR code on each card opens this lesson.

More cards: Ravi Sir's Cards.