Skip to content

awk_cmd.md

Latest commit

 

History

History
45 lines (41 loc) · 2.31 KB

awk_cmd.md

File metadata and controls

45 lines (41 loc) · 2.31 KB

AWK Command

The awk command is a general purpose stream processing application that is similar to SED. However, awk can be used to generate much larger programs and is better at handling columnar data that is space separated.

This tutorial is not an exhaustive user guide for awk. Only some useful tricks and tips that apply to data manipulation will be covered here.

General Form

awk [-F _sep_] 'BEGIN {_init_action_} _pattern_ {_action_} ... END {_end_action_}'
  • _sep_ : an optional field separator can be specified as flag or using FS variable in BEGIN block.
  • _init_action_ : code specified inside the optional BEGIN special pattern is executed before any of the pattern matching occurs. This is usually used to initialize counter variables, or print header columns.
  • _end_action_ : code specified inside the optional END special pattern is executed after all the pattern matching actions are performed. This is usually used to print summary variables or table footers.
  • _pattern_ : a mathematical or regular expression. The associated _action_ is only executed for lines where _pattern_ is true. The awk program can contain one or more _pattern_/_action_ pairs.

Special Variables

The following variables are available inside the action blocks.

  • NR : number of records seen so far. Starts at 1.
  • NF : number of fields in current record.
  • FS : current input field separator (can be set using -F flag)

Examples

Print columns 3 and 1 (in that order) from rows 2 to 4. The pattern 2<=NR && NR<=4 will match rows 2 to 4 and only on these rows, will the print occur. If the pattern was left blank, then the selected columns from all rows from the input will be shown.

$ awk -F: '2<=NR && NR<=4 {print $3"|"$1}' input_users.txt
David|3
Carol|5
Carol|11

print rows where the ACCT_NUM column (column #2) is even. The expression $2 % 2 will be 0 for even values and 1 for odd values of column 2.

$ awk -F: '$2%2 == 0 {print}' input_users.txt
USER_ID:ACCT_NUM:USER_NAME
1:1234:Frank
7:9876:Alice
4:666:Bob
11:1234:Carol
4:5566:Bob

Compute sum of ID column. The column header "USER_ID" has a value of 0 when converted to numeric.

$ awk -F: 'BEGIN {SUM=0} {SUM+=$1} END {print "Sum of ID=",SUM}' input_users.txt
Sum of ID= 55