Multiple field sort (quick explanation)

bioinformatics_guy · 01-12-2009, 07:28 AM

Lets say I had a file of the following form:

a 0 0
a 0 1
a 1 0
a 1 1
b 0 0
b 0 1
b 1 0
b 1 1

And I want to sort first by the 2nd column and then by the first column.

I know that the command is:

sort +1 -2 +0 -1 <filename>

but my question is, I know the +1 means sort by the second column but what does the -2 mean? Same with +0 -1. Why can't I just put:

sort +1 +0 <filename>

Which I would expect to sort by the second column and if thats the same sort by the first; which is not correct.

onebuck · 01-12-2009, 07:31 AM

Hi,

Smells like homework to me!

You could 'man sort'.

pixellany · 01-12-2009, 07:55 AM

Where are you getting that syntax? It does not appear in the man page.

bioinformatics_guy · 01-12-2009, 09:09 AM

I found it on some Unix help documentation online. Apparently, it is the old syntax. The equivalent is:

sort -k2,2 -k1,1 <Filename>

which seems to make more sense as the index starts at 1, the first number is what field to start and the second is which field to stop at (include that field).