[SOLVED] Bash "sort" command problem

rt1870 · 04-12-2013, 08:13 AM

I am trying to use a piece of scientific software that involves executing a number of scripts and commands.There is a particular step I am stuck on, where the program needs a file to be sorted and the documentation gives the command:

sort +3 -25 file1 >file2

The options +3 and -25 do not seem to correspond to any of the options for "sort" when I look at the help in bash. Am I missing something simple here? Is this not actually a bash command and something else? Unfortunately the documentation does not explain the way in which the data needs to be sorted, so I am unable to implement my own solution.

I am fairly new to using bash and linux in general and am completely stuck. Any help appreciated!

bigrigdriver · 04-12-2013, 09:56 AM

Quote:

I am trying to use a piece of scientific software that involves executing a number of scripts and commands.There is a particular step I am stuck on, where the program needs a file to be sorted and the documentation gives the command:

sort +3 -25 file1 >file2

It would help if you identified that "scientific software" so that others can examine the documentation.

rknichols · 04-12-2013, 10:09 AM

sort is a separate program and not something internal to bash. The command line you were given uses an obsolete syntax, and specifies a sort beginning with field 3 and extending up to but not including field 25, where the fields are numbered from 0. The current manual page for sort no longer describes this syntax, though the sort command itself does still accept it. You can find a page that includes the obsolete syntax here.

The modern equivalent would be

Code:

sort -k4,25 file1 >file2

The field numbers here start from 1, and the sort is inclusive of the starting and ending field numbers.

konsolebox · 04-12-2013, 10:12 AM

I think they are key field numbers and are somehow similar to the values presented with -k only that the position is n + 1. e.g. sort +3 -25 is similar to sort -k4,25. The numbers presented in -k represent the key fields of the line which are prioritized in sorting before the whole line is sorted in general.

So if I have a file containing

Code:

aaa	def	ghi
bbb	abc	ghi
aaa	abc	jkl

Code:

sort +1 -3 file.txt
sort -k 2,3 file.txt
sort -k 2 file.txt

will have

Code:

bbb     abc     ghi
aaa     abc     jkl
aaa     def     ghi

and

Code:

sort +1 -2 file.txt
sort -k 2,2 file.txt

will have

Code:

aaa     abc     jkl
bbb     abc     ghi
aaa     def     ghi

David the H. · 04-14-2013, 08:31 AM

Try reading "info sort". It describes the usage of the program in a lot more detail than the manpage.

This is true of most info pages, by the way.

rt1870 · 04-17-2013, 07:12 AM

Just a bump to say thank you to the replies - answered my question perfectly (especially konsolebox!).