UNIX for Dummies Questions & Answers

Hi all..,
I have problem when i want to take every number in my file.

this is my file look like..

errorpath.csv

Code:

error=29
error=34
error=2
error=3
error=6
error=5
error=36
error=36
error=36
error=36
error=36
error=36
error=36
error=36
error=36
error=36

from this file i want to take every different number.

Actually that file was generated by myprogram. I can make the result just a number.
for example

errorPath.csv

Code:

32
32
4
4
4
4
36
2
2
2
2
2
2
2
2
2
2

I think I can solve my task if I can make new file by get data from errorPath.csv by make other file like this.
this is file that i expect can maked.

result.cfgvar1=32
var2=4
var3=36
var4=2
I need shell script to that task and I think that SED command can help me.
Thanks before..

That is a strange name for a file that contains no commas, but, assuming your sample data is representative of your real data, the following should do what you want:

Code:

awk '!A[$0]++' errorpath.csv

If you want to try this on a Solaris/SunOS system, change awk to /usr/xpg4/bin/awk.

If it's just for the number, modify Don Cragun's proposal slightly:

Code:

awk '!A[$2]++ {print $2}' FS="=" errorpath.csv

Can I do it by SED command?

I'm afraid, no. sed doesn't have variables to keep all the duplicate values.

I don't know if you can do it or not. Editors (including ed, ex, and sed) are not designed to perform logical evaluations like this. I can imagine using one sed script to extract the numbers found in your file and using a shell while read loop to feed each of those numbers to another sed script that could then be used delete duplicate occurrences of that number. But, invoking sed N + 1 times for an N line file (instead of invoking awk once) would be extremely inefficient (and would not be just using sed.

I think nice idea if I can take the value and put its value to a variabel in one other file.
But totally I dont know how to make it real..