Sort question

[pjb]

I have a multiple field file that I need to sort by field 1, and then by field 3 within field 1. I also have to drop the duplicates of field 1, keeping only one occurrence of field on that has the lowest value of field 2.
Can this be done with one or more executions of sort?

 

[gyan]

Hi Pjb,

Yes, you can!

Use sort and uniq.

Thanks,

Gyan

 

[pjb]

Yes, but the -u parameter will drop duplicates. These are not duplicates. In field 1 the value is the same, but in field 2, the values can differ. The file would look like this:

AAAA 1
AAAA 2
BBBB 1
BBBB 2

How would I retain only the AAAA and BBBB records that had the lowest value in field 2?

 

[butterfm]

This is probably not the best solution but it should work if i'm understanding correctly what you're asking.

for i in `cat testfile | awk '{print $1}' | sort -u`
do
grep $i testfile | sort +0 -1 -t" " +1 -n | head -1
done

 

[bigoldbulldog]

Assuming this is all sorted and your input is

AAAA 1
AAAA 2
BBBB 1
BBBB 2

then try

awk '{if($1!=FLD) print; FLD=$1}' input-file

which only keeps the first record based on the first field.

Cheers,
ND

 

[aigles]

Hi pjb,

Try :

sort -k1,3 input_file | sort -u -k1,1 Jean Pierre.