topic Re: question with join command.. in Operating System - HP-UX

question with join command..

amonamon — Mon, 18 Dec 2006 09:49:53 GMT

Hello..I have 2 files:

file1(pattern):

100018
100090
100143
100198
100346
100708
100961
100989
101037
101206
101543
101931
102306
102371
102504
102605
102640
102823
103025
103390

file2:

320346|006|12|11
901546|007|1|11
220054|007|12|11
100018|007|1|11
901411|007|12|11
380094|007|12|11
320046|006|12|11
901946|007|1|11
220854|007|12|11
310776|007|1|11
901611|007|12|11
380194|007|12|11

result should be:
100018

so basicly join 2 files and if record is listed in pattern(file1) do not print that record..

this is what I tryed so far:

join -t'|' -a1 file1 file2

but results are not OK..

then I also tryed this:

join -1 1 -2 1 -t'|' -a 1 file1 file2

still not OK??

can someone correct my misstake..

Thanks a lot..I am still beginner here but I try hard to make something..

cheers,

Re: question with join command..

spex — Mon, 18 Dec 2006 09:56:02 GMT

Hi,

From your example, it's not clear if you DO or DO NOT wish to print matching records in file2.

If you DO want matching records printed:
$ grep -F -f file1 file2
100018|007|1|11

If you wish to EXCLUDE matches in file2:
$ grep -F -v -f file1 file2
320346|006|12|11
901546|007|1|11
220054|007|12|11
901411|007|12|11
380094|007|12|11
320046|006|12|11
901946|007|1|11
220854|007|12|11
310776|007|1|11
901611|007|12|11
380194|007|12|11

PCS

Re: question with join command..

James R. Ferguson — Mon, 18 Dec 2006 10:03:36 GMT

Hi:

If you truly want to use 'join' you need to sort your input files. See the manpages for 'join' for more information.

Regards!

...JRF...

Re: question with join command..

amonamon — Mon, 18 Dec 2006 10:07:51 GMT

Yes I learned man join and I realised and invoked sort -b file1

It works with command:

grep -F -f file1 file2

cool..but still I wanted to practice join command becouse I find it cool..

Thanks a lot

Re: question with join command..

Sandman! — Mon, 18 Dec 2006 10:41:33 GMT

The grep command is perfect for what you're trying to do. But if you wanted to practice the join(1) command you need to read the man page thoroughly as the second paragraph under DESCRIPTION section of the join(1) man page says:

>file1 and file2 must be sorted in increasing collating sequence (see >Environment Variables below) on the fields on which they are to be joined; >normally the first in each line.

So sort file1 and file2 before using the join(1) command on them, i.e.

# sort -k1,1 file1 > file1.out
# sort -t'|' -k1,1 file2 > file2.out
# join -1 1 -2 1 -t'|' -o 1.1 file1.out file2.out

~cheers