[SOLVED] awk

danielbmartin · 05-17-2012, 03:04 PM

Some months ago I asked for an awk which would sort all words within each line.
firstfire provided this excellent solution using asort.

Code:

awk -v IGNORECASE=1 '{gsub(/[[:punct:]]/, ""); split($0, w); s=""; for(i=1; i<=asort(w); i++) s=s w[i] " "; print s }' infile.txt

Now I want to do something similar but have been unable to modify it to do the job. I have a file with one character string per line. The characters in this example file are numerics but the question applies to alphanumerics.

Have:

Code:

Want:

Code:

Daniel B. Martin

Kustom42 · 05-17-2012, 03:15 PM

the sort command will do that for you by itself.

Kustom42 · 05-17-2012, 03:17 PM

Wait, after looking at your output there is no real sorting method there, how are you trying to sort these?

colucix · 05-17-2012, 03:17 PM

I suggest something similar but simpler:

Code:

awk '{ split($0,array,""); for ( i = 1; i <= asort(array); i++ ) printf array[i]; printf "\n" }' file

The sorting process depends on your locale definitions and specifically from LC_COLLATE. You can use IGNORECASE as well.

danielbmartin · 05-17-2012, 03:40 PM

Quote:

Originally Posted by colucix

Code:

awk '{ split($0,array,""); for ( i = 1; i <= asort(array); i++ ) printf array[i]; printf "\n" }' file

Ooh, that is sweet! Thank you!

Daniel B. Martin

grail · 05-18-2012, 09:39 AM

And a bit of ruby:

Code:

ruby -ane 'BEGIN{$; = ""};puts $F[0..-2].sort.join' file

or

ruby -pe '$_ = $_.chomp.split("").sort.join + "\n"' file

jachaja · 11-07-2012, 07:20 PM

Hi all,

I have a file with a lot of records similar to this:

023456AB2543999210709
022656AB2543999210409
024556AB2543999200709
024556AB2543999200909

I want to sort this file by using two parts of each record, for example: positions 3 to 4 like one field ascending and
positions 16 to 19 like other field descending.

The result:
022656AB2543999210409
023456AB2543999210709
024556AB2543999200909
024556AB2543999200709

It is posible?

Thanks.

danielbmartin · 11-07-2012, 08:52 PM

Quote:

Originally Posted by jachaja

Hi all,

I have a file with a lot of records similar to this:

023456AB2543999210709
022656AB2543999210409
024556AB2543999200709
024556AB2543999200909

I want to sort this file by using two parts of each record, for example: positions 3 to 4 like one field ascending and
positions 16 to 19 like other field descending.

The result:
022656AB2543999210409
023456AB2543999210709
024556AB2543999200909
024556AB2543999200709

It is posible?

Thanks.

Try this ...

Code:

echo "Sort on columns 3-4 (ascending) and columns 16-19 (descending)."
sort -k1.3,1.4 -k1.16,1.19r $InFile > $OutFile

Daniel B. Martin