Shell Programming and Scripting

Hi all,

I have .ksh script, which is doing different things like creating output files, renaming them, sending them via ftp and save a backup of the files on the application server. Now I'd like to edit this file to be able to split the incoming file into 4 different files. Therefor the database pkg creating the content of the file adds prefixes to every single line. So the file looks like:

BDDandsomedifferentinformation
BDTandalsosomeinformation
BNDandsomeotherinformation
BADandagainsomeinformation

Tried to manually split that file with awk and it works fine:

Code:

awk '{close(f);f=$1}{print > f".txt"}' BHD000000744.TXT.3292214

But when I want to have this in the .ksh, I do not get any result?! It looks like it is doing nothing. Tried it that way:

Code:

if [[ ${_ftpfile} = @(BHD*) ]]; then
  awk '{close(f);f=$1}{print > f".txt"}' $_ftpfile
fi

$_ftpfile is the filename of the file which should be split, and the if checks if the filename starts with BHD, which also works, but as i said, not within the existing .ksh.

Any idea what the reason for that can be? Appreciate any help, thanks!

Mike

put "set -x" before the if statement and post the output please.

Code:

set -x

if [[ ${_ftpfile} = @(BHD*) ]]; then
  awk '{close(f);f=$1}{print > f".txt"}' $_ftpfile
fi

set +x

Sorry, but set -x does what?
The loop is in a .ksh script which is called from out an application with different parameters, and there is e.g. a file (variable $_dbgfile) which is used, to log the output.

I'm pretty new in this, so thanks for your patience

---------- Post updated at 11:14 AM ---------- Previous update was at 09:09 AM ----------

ok, figured it out, so here's the output:

+ [[ BHD000000813.TXT = @(BHD*) ]]
+ awk {close(f);f=$1}{print > f".txt"} BHD000000813.TXT
+ echo Split file in
+ 1>> ./../log/CSW_OUT_ORDERS.3292291.dbg
+ ./awk.ksh BHD000000813.TXT


I've tried to call awk in a seperate .ksh file, as it did not work, so thats why there's s seberate line at the end for that.

Code:

set -x
if [[ ${_ftpfile} = @(BHD*) ]]; then
  awk '{close(f);f=$1}{print > f".txt"}' $_ftpfile
  #/appl/spider/if/SME/csw/script/awk.ksh $_ftpfile
  echo "Split file in " >> $_dbgfile
  ./awk.ksh $_ftpfile
fi
set +x

The only thing I see different in your original post and the output you just posted is the filenames:

original: BHD000000744.TXT.3292214
output: BHD000000813.TXT

Can you put a sample of the file you want parsed on here and show what files you want it broken into and what the names and contents of each should look like as they relate to the sample file?

Thanks.

The filename is different, cause I tried to awk manually on a file which was already transferred via ftp, which we mark with a suffix of the process id then.

Data is just dummy date, and files really look like:

BDD000000821 00000681 RD00000681 103000 1 01934110012-63 TEST2 11-FEB-110 10 1 000 SHIPPING
BDD000000821 00000681 RD00000681 103000 1 01934110012-63 TEST2 11-FEB-110 10 1 000 SHIPPING
BND000000821 000006831916000006831916BILL OF REMARKSALLOVERANDOVER
BND000000821 000006831918000006831918BILL OF LAADSFASDFASFDASDF#ADFADSF
BAD000000821 00000681 00000681 103000 BESTBEST #123 1234 SOMEWHERE
BAD000000821 00000681 00000681 419000 ALLTBEST #456 1234 SOMEWHERE

Funny thing is, that if I start the ksh script manually, which calls a script awk.ksh (see below) to do the split, then it works great. Starting the process from out the application, does nothing ...

Code:

#!/usr/bin/ksh
_file_name=$1
awk '{close(f);f=$1}{print > f".txt"}' $_file_name
echo "awk has been called" >> tmp.log

---------- Post updated at 03:41 PM ---------- Previous update was at 01:58 PM ----------

The problem is sovled now ... due to the startup in the application, it was simply a completely different working directory which I did not check

Files are there, now the next question:

Is there any way, to put the files from the awk into variables in the ksh?

Code:

awk '{close(f);f=$1}{print > f".txt"}' $_ftpfile

means, the .txt files, as many as there are generated, into file1, file2, ... variables

You can do the job with one command IMHO:

Code:

awk '{close(f);f=$1}{print > f".txt"}' BHD*

Maybe some missunderstanding ... $_ftpfile is the file which is split into 2 or more files, and print > f".txt" is writing the filenames, right?

So I want this filenams in variables, so that finally $_ftpfile has been split and is in variables $_ftpfile1, $_ftpfile2, ...