Overview
AWK is a common command in Linux, there are GAWK, NAWK and OAWK, generally Linux distributions use GAWK.
AWK handles text in rows, and by default uses space or \t as a separator to group a line and fill the corresponding field with the specified variable. The command takes the form of
awk 'condition{execute statement} condition{execute statement}' filename
There are two input sources
- The file specified by the command
- The output of the pipe
Multiple conditions are independent of each other, and will match the conditions one by one, and execute the response execution statement if the condition is met
Built-in variables
- $0 for a whole line
- $N is the Nth field
- NF number of fields per line
- NR the number of rows the line is currently on
- FS current split character, default is space bar or Tab
- OFS Output field split character, defaults to space
- RS record split character, defaults to newline
- ORS record separator for output, defaults to newline
- FILENAME The name of the current input file
Examples of using conditional and execution statements.
awk '{FS=":"} $3<100{print $1 "\t" $3}' /etc/passwd
The output is.
The first line takes effect
The first line does not take effect, to make the first line take effect you need to use the keyword BEGIN
awk 'BEGIN {FS=":"} $3<100{print $1 "\t" $3}' /etc/passwd
Multiple commands
awk 'BEGIN {FS=":"} {print $1; print $2}' /etc/passwd
Multiple commands can exist in command brackets, but they should be separated by a “;” sign, or naturally separated by a line break.
String matching
Actually AWK can be used instead of grep, for example, our normal grep: !
If you change it to awk, it would be
Then there’s no grep, what? grep -v works well? Try this one.
awk '! /root/ { print $1}' /etc/passwd
Field Filtering
The example just demonstrated can only filter one line, however, we can do much more, we can filter only the username: ```
awk 'BEGIN {FS=":"} $1 ~ /root/{print $1}' /etc/passwd
BEGIN and END
- BEGIN { action before subsequent command execution }
- END { action after command execution }
Environment variables
Environment variables can be used in AWK as the variable ENVIRON, either by exporting the environment variable or by passing the -v argument when calling awk
$ awk -v val=$x '{print $1, $2, $3, val, ENVIRON["ts"]}' OFS="\t" /etc/passwd