....How to use awk to sort....

nisson · 10-15-2004, 06:09 PM

How to use awk command in linux to sort
a class of float-type datas, which is saved as
in a file. Thanks !

homey · 10-15-2004, 06:11 PM

Do you have an example of what you want to sort?

nisson · 10-15-2004, 09:34 PM

For example, I have a file 2.dat, whch is shown as follows.
0 0 -0.000000
0 2 -0.067044
0 3 -2.114625
0 4 -3.520602
0 5 -5.153252
0 6 -3.830406
0 7 -5.566137
0 8 -7.199962
0 9 -8.227023
0 10 -6.939604
0 11 -8.229088
0 12 -9.169961
0 13 -9.821586
0 14 -8.801335
0 15 -9.677837
0 16 -10.200804
0 17 -9.708465
0 18 -9.827853
0 19 -9.895395
0 20 -9.746067
0 21 -9.560159
0 22 -9.148484

I want to produce a file 2.dat which is sorted according to the value of $3.

Thanks

homey · 10-15-2004, 10:08 PM

I'm not sure which way you want it to go. Try the sort command with or without -r
cat /home/file.txt | awk '{print $3}' | sort -nr > /home/file1.txt

nisson · 10-15-2004, 10:33 PM

Thanks! How to add $1, $2 to the sorted file
since the sorted file just contain the $3 according to
your given command.

nisson · 10-15-2004, 10:41 PM

Thanks! I have find a combined comands to
add $1,$2 to the file.

homey · 10-16-2004, 04:05 AM

Ok that was a fun challenge!

I never did figure it out but found a web site with the sollution. http://www.linuxfocus.org/English/Se...rticle103.html .
It's down the page in the problem / solution section.

Getting the program to sort $3 is easy enough but the trick is getting $1 and $2 into the output. With the sollution from that web site, $3 ends up in front of the others which I figured you wouldn't want so I added some more awk statements to get them back where they belong.

Code:

#!/bin/bash

stuff="/home/file.txt"
cat ${stuff} | \
while read num
do
awk '{printf $NF;$NF = " "; printf " "$0"\n" }' | sort -nr | \
awk '{printf $NF;$NF = " "; printf " "$0"\n" }' | \
awk '{printf $NF;$NF = " "; printf " "$0"\n" }' > file1.txt

done

bigrigdriver · 10-16-2004, 06:55 PM

awk 2.dat ' $ {print $0 }' | sort -k 3 -n -r -o 2.dat 2.dat

This will read file 2.dat, print the entire record, sort on field 3, print in numerical order (reversed), output to 2.dat, read from 2.dat.
It generates an error message (I haven't figured out why yet) about not being able to open file 2.dat, but it also sorts 2.dat, from largest to smallest value. To sort smallest to largest, remove -r from the above command.
So. Error message or no error message, the file gets sorted.

PS. The sort works fine without getting awk involved.

homey · 10-16-2004, 07:42 PM

Yep, you're right and that's what I get for staying up too late.

Code:

awk '{print $0}' | sort -k 3 -nr file.txt > file1.txt

or even this...

sort -k 3 -nr file.txt > file1.txt

nisson · 10-16-2004, 10:15 PM

Thank both of you!

homey · 10-16-2004, 11:30 PM

You're welcome from here!

What did your solution look like?

nisson · 10-17-2004, 10:04 PM

cat /home/file.txt | awk '{print $3,$1,$2}' | sort -nr > /home/file1.txt
awk '{print $2,$3,$1}' file1.txt >file2.txt

"sort -k 3 -nr file.txt > file1.txt" is more simple and good way to do it.

vipul_makwana · 11-02-2012, 03:51 PM

Quote:

Originally Posted by bigrigdriver

awk 2.dat ' $ {print $0 }' | sort -k 3 -n -r -o 2.dat 2.dat

This will read file 2.dat, print the entire record, sort on field 3, print in numerical order (reversed), output to 2.dat, read from 2.dat.
It generates an error message (I haven't figured out why yet) about not being able to open file 2.dat, but it also sorts 2.dat, from largest to smallest value. To sort smallest to largest, remove -r from the above command.
So. Error message or no error message, the file gets sorted.

PS. The sort works fine without getting awk involved.

First of all thanks for the answer it helped me.....but If you will change the command a bit than the Error message wont come. I have tried it myself.
The correct syntax would be.........
$ awk '{print $0 }' 2.dat | sort -t "|" -k 6 -n -r -o 2.dat 2.dat

that means don't put "$" in front of print option..............