UNIX for Dummies Questions & Answers

Hello!
I have a file with 4 columns. I am trying to have it sort first with respect to the first column, and then with respect to the number of counts (in descending count) in the second column within the same first column identity.

For example:

Input:
1 2 A 1
1 6 B 2
2 5 G 7
1 6 D 4
1 6 F 6
1 0 E 5
2 4 H 8
2 5 I 9
1 2 C 3

The output is to look like:
1 6 B 2
1 6 D 4
1 6 F 6
1 2 A 1
1 2 C 3
1 0 E 5
2 5 G 7
2 5 I 9
2 4 H 8

As you can see:
-Within the first column identity=1, there are 3 counts of '6', 2 counts of '2' and 1 count of '0' in the second column.
-Within the first column identity=2, there are 2 counts of '5' and 1 count of '4' in the second column.
They are both sorted with respect to descending count in the second column.

I have been doing this in excel and it became quite time consuming for me since the number of files to process has started to increase.

Thanks for your help in advance!

Code:

sort -n -k1 -k2 inputfile > outputfile

The sort in the second column is not (necessarily) in arithmetic order, it is rather based on the reverse count.

Thanks!

You changed your output sample

Code:

#!/bin/ksh

 rm *.tmp
 sort -n -k1 filename |  \
   awk '{file=sprintf("%d.tmp", $1); print $0 > file}'
 
 for fname in *.tmp
 do 
    sort -n -r -k2 $fname
 done > outputfile
 rm *.tmp

Note - do this in a small subdirectory it uses tempfiles. If your first column has too many distinct values it will not work either. This depends on system's your open file limit. You'll have to resort to something else.

Code:

getconf OPEN_MAX

tells you how many open files you can have at one time.

Sorry; I just wanted to make it more clear... (I guess I could not succeed =( )

Thanks for the code... However, it still sorts the second column in arithmetic order, not with respect to the number of counts each item has within each group...

I hope the following makes it a little more clear:
Input:
1 2 A 1
1 1 B 2
2 4 G 7
1 1 D 4
1 1 F 6
1 0 E 5
2 5 H 8
2 4 I 9
1 2 C 3

The output is to look like:
1 1 B 2
1 1 D 4
1 1 F 6
1 2 A 1
1 2 C 3
1 0 E 5
2 4 G 7
2 4 I 9
2 5 H 8

Thanks again!

Code:

1 1 B 2
1 1 D 4
1 1 F 6
1 2 A 1
1 2 C 3
1 0 E 5
2 4 G 7
2 4 I 9
2 5 H 8

This is not in any order I understand - with regard to column #2

1) Within the first column identifier = 1:
There are *3* counts of '1',
*2* counts of '2',
*1* count of '0'

2) Within the first column identifier = 2:
There are *2* counts of '4',
*1* count of '5'

They are ordered with respect to the number of counts of each item within the group, starting with the item with the most count to the least, independent of their values...