Sort lines by group and column

Question

I have a csv file separated by semicolons. Which contains lines as shown below.. And I need to sort it by the first and third column, respecting the groups of lines defined by the value of the first column.

booke;book;2
booke;booke;1
booke;bookede;6
booke;bookedes;8
booke;booker;4
booke;bookes;7
booke;booket;3
booking;booking;1
booking;bookingen;2
booking;bookingens;3
booking;bookinger;7
booking;bookingerne;5
booking;bookingernes;6
booking;bookingers;8
booking;bookings;4

Expected output:

booke;booke;1
booke;book;2
booke;booket;3
booke;booker;4
booke;bookede;6
booke;bookes;7
booke;bookedes;8
booking;booking;1
booking;bookingen;2
booking;bookingens;3
booking;bookings;4
booking;bookingerne;5
booking;bookingernes;6
booking;bookinger;7
booking;bookingers;8

I tried it with sort -t; -k3,3n -k1,1 but it's sorted by the third entire column.

Switch `-k3,3n` and `-k1,1`? – Benjamin W. Jan 03 '16 at 21:03 — Benjamin W., Jan 03 '16 at 21:03

score 2 · Accepted Answer · edited May 23 '17 at 12:15

What about using two sorts in a pipeline fashion:

sort -t ';' -k 3,3n | sort -t ';' -k 1,1 -s

The -s in the second parameter is necessary in order to enable stable sort. Otherwise it could destroy the previous (third column) sorting.

EDIT: however as @BenjaminW. points out in his comment, you can use multiple -k flags, you only specified them the wrong way. By performing a sort:

sort -t ';' -k 1,1 -k 3,3n

It takes the first column als primary sorting column and the third as secondary.

Sort lines by group and column

1 Answers1

Linked