← Shell / Bash Course5 / 13

Redirection & Pipelines

Control where a command reads from and writes to: redirect stdout and stderr, append to files, discard output, chain commands with pipes, and capture or stream output with command and process substitution.

Ad 728×90

The three standard streams

Why: every command has three channels — standard input (0), standard output (1), and standard error (2). Normal results go to stdout; error messages go to stderr. Keeping them separate lets you save output while still seeing errors, or the reverse.

# stdin  (0) — where a command reads input
# stdout (1) — normal output
# stderr (2) — error and diagnostic output

ls /etc /does-not-exist
#   /etc listing  -> stdout
#   "No such file" error -> stderr

Redirect output to a file

Why: > sends stdout to a file, replacing its contents. >> appends instead of overwriting. To redirect errors, target stream 2 with 2>. To send both to one place, use &> (or 2>&1).

Overwrite a file with the output

echo "first line" > log.txt

Append another line

echo "second line" >> log.txt

Send errors to their own file

ls /nope 2> errors.txt

Send BOTH stdout and stderr to one file

ls /etc /nope &> all.txt

Discard output with /dev/null

Why: /dev/null is a "black hole" — anything written to it disappears. Redirect there when you only care whether a command succeeded, not what it printed. 2>/dev/null hides errors; >/dev/null 2>&1 hides everything.

Silence normal output, keep errors visible

command --version > /dev/null

Silence errors only

grep "x" file.txt 2> /dev/null

Silence everything — just check the exit code

ping -c1 example.com > /dev/null 2>&1

Pipes: chain commands together

Why: a pipe | feeds one command's stdout into the next command's stdin, so you build a pipeline that transforms data step by step. Note: the text tools in these pipelines (grep, sort, uniq, wc, cut) are covered in the Linux course lesson on pipes & text processing.

Count how many lines contain "error"

cat app.log | grep "error" | wc -l

Top 3 most common values in a column

cut -d',' -f2 data.csv | sort | uniq -c | sort -rn | head -3

Process substitution

Why: <(command) makes a command's output look like a temporary file, so you can feed two command outputs into a tool that expects file arguments — like diff comparing two live listings. This is a Bash-specific feature plain sh does not have.

Compare the output of two commands without temp files

diff <(ls dir-a) <(ls dir-b)

Feed a command's output where a file path is expected

while read -r line; do

echo "got: $line"

done < <(grep "WARN" app.log)