Awk Command Practice
awk splits each record into fields ($1…$NF) and runs a pattern-action program. This track drills printing columns, /pattern/ and NR filters, -F separators, and BEGIN/END sums — the one-liners you type in a pipeline, not a spreadsheet. Grep selects lines; sed edits them; awk computes on fields.
Command Reference
separator
| Command | Description |
|---|---|
awk -F: '{print $1}' <file> | Split each record on colon and print the first field. |
awk -F: '{print $1, $NF}' <file> | Split each record on colon and print the first and last fields. |
awk -F ',' '{print $NF}' <file> | Split each record on comma and print the last field. |
awk 'BEGIN {FS=":"; OFS=","} {print $1,$2,$3}' <file> | Read colon fields and print the first three fields separated by commas. |
awk -v FS=':' '{print $1}' <file> | Set FS with -v and print the first field (not the -F spelling). |
| Command | Description |
|---|---|
awk '{print $1, $NF}' <file> | Print the first field and the last field of every record. |
awk '{print NR, $0}' <file> | Print the record number and the whole line for every record. |
awk '{print $1}' <file> | Print the first field of every record. |
awk '{print $(NF-1)}' <file> | Print the second-to-last field of every record. |
awk '{print $1 "=" $2}' <file> | Print the first two fields joined by an equals sign, with no extra OFS space. |
awk '{$1 = NR; print}' <file> | Replace the first field with the record number and reprint the record. |
select
| Command | Description |
|---|---|
awk 'NR > <num>' <file> | Print records after a given record number. |
awk '$NF > <num>' <file> | Print records whose last field is greater than a number. |
awk 'NR==<num>,NR==<num>' <file> | Print an inclusive range of record numbers. |
awk 'NR%<num>==1' <file> | Print every Nth record, starting with the first. |
script
| Command | Description |
|---|---|
awk -f <script> <file> | Run commands from an awk script file against an input file. |
awk -F: -f <script> <file> | Run a script file with colon as the field separator. |
awk -v <name>=<text> '{print <name>}' <file> | Pass a variable into the program and print it on every record. |
awk -v OFS=',' '{print $1,$2}' <file> | Set OFS with -v and print the first two fields comma-separated. |
blocks
| Command | Description |
|---|---|
awk 'BEGIN {print "<text>"}' | Print a string once before reading any input. |
awk 'BEGIN {print "<text>"} {print $1}' <file> | Print a header once, then the first field of every record. |
awk '{print $1} END {print "<text>"}' <file> | Print the first field of every record, then a trailer once. |
awk '{s+=$1} END {print s}' <file> | Add the first field of every record and print the total once. |
filter
| Command | Description |
|---|---|
awk '/<pattern>/' <file> | Print records whose whole line matches a pattern. |
awk '/<pattern>/ {print $2}' <file> | Print the second field of records whose line matches a pattern. |
awk '!/<pattern>/' <file> | Print records whose whole line does not match a pattern. |
awk '$1 ~ /<pattern>/' <file> | Print records whose first field matches a regex. |
awk '$1 !~ /<pattern>/' <file> | Print records whose first field does not match a regex. |
awk '$1 == "<name>"' <file> | Print records whose first field equals an exact string. |
Key Use Cases
- Print the first and last colon-separated fields from a passwd-style file
- Keep only lines whose first field matches a regex
- Number every record and print the whole line
- Sum the first column and print the total in an END block
- Set FS and OFS so comma output matches colon input
- Run a saved .awk script file against an input file
Frequently Asked Questions
How is awk different from sed?
sed applies an address plus an edit (s///, d, i). awk splits the line into fields and runs a program. print $1 and END {print s} are awk; they are not sed scripts.
What are $0, $1, and $NF?
$0 is the whole record. $1 is the first field. $NF is the last field; NF is the field count, so $(NF-1) is the second-to-last field.
When do I use -F versus BEGIN {FS=…}?
-F (or -v FS=) sets the input field separator before records are read. BEGIN {FS=":"; OFS=","} is the same idea when you also need OFS or other setup. This track does not treat -F and -v FS= as fill-blank aliases of each other.
What do BEGIN and END run on?
BEGIN runs once before any input. END runs once after the last record. A missing pattern (just {…}) runs on every record.
Ready to master Awk commands?
Test your muscle memory with our spaced-repetition quiz system. Free forever.