rearrange columns using awk or cut command

Question

I have large file with 1000 columns. I want to rearrange so that last column should be the 3rd column. FOr this i have used,

cut -f1-2,1000,3- file > out.         


        

Answer 1

I THINK what you want is:

awk 'BEGIN{FS=OFS="\t"} {$3=$NF OFS $3; sub(OFS "[^" OFS "]*$","")}1' file

This might also work for you depending on your awk version:

awk 'BEGIN{FS=OFS="\t"} {$3=$NF OFS $3; NF--}1' file

Without the part after the semi-colon you'll have trailing tabs in your output.

Answer 2

try this awk one-liner:

awk '{$3=$NF OFS $3;$NF=""}7' file

this is moving the last col to the 3rd col. if you have 1000, then it does it with 1000th col.

EDIT

if the file is tab-delimited, you could try:

awk -F'\t' -v OFS="\t" '{$3=$NF OFS $3;$NF=""}7' file

EDIT2

add an example:

kent$  seq 20|paste -s -d'\t'                              
1   2   3   4   5   6   7   8   9   10  11  12  13  14  15  16  17  18  19  20

kent$  seq 20|paste -s -d'\t'|awk -F'\t' -v OFS="\t" '{$3=$NF OFS $3;$NF=""}7'
1   2   20  3   4   5   6   7   8   9   10  11  12  13  14  15  16  17  18  19  

EDIT3

You didn't give any input example. so assume you don't have empty columns in original file. (no continuous multi-tabs):

kent$  seq 20|paste -s -d'\t'|awk -F'\t'  -v OFS="\t" '{$3=$10 FS $11 FS $3;$10=$11="";gsub(/\t+/,"\t")}7'
1       2       10      11      3       4       5       6       7       8       9       12      13      14      15      16      17      18      19      20

After all we could print those fields in a loop.

Answer 3

Since many people are searching for this and even the best awk solution is not really pretty and easy to use I wanted to post my solution (mycut) written in Python:

#!/usr/bin/env python3

import sys
from signal import signal, SIGPIPE, SIG_DFL
signal(SIGPIPE,SIG_DFL)

#example usage: cat file | mycut 3 2 1

columns = [int(x) for x in sys.argv[1:]]
delimiter = "\t"

for line in sys.stdin:
    parts = line.split(delimiter)

    print("\t".join([parts[col] for col in columns]))

I think about adding the other features of cut like changing the delimiter and a feature to use a * to print the remaning columns. But then it will get an own page.

Answer 4

A shell wrapper function for awk' that uses simpler syntax:

# Usage: rearrange int_n [int_o int_p ... ] < file
rearrange () 
{ 
    unset n;
    n="{ print ";
    while [ "$1" ]; do
        n="$n\$$1\" \" ";
        shift;
    done;
    n="$n }";
    awk "$n" | grep '\w'
}

Examples...

echo foo bar baz | rearrange 2 3 1
bar baz foo 

Using bash brace expansion, rearrange first and last 5 items in descending order:

echo {1..1000}a | tr '\n' ' ' | rearrange {1000..995} {5..1}
1000a 999a 998a 997a 996a 995a 5a 4a 3a 2a 1a 

Sorted 3-letter shells in /bin:

ls -lLSr /bin/?sh | rearrange 5 9 
150792 /bin/csh 
154072 /bin/ash 
771552 /bin/zsh 
1554072 /bin/ksh