topic Re: awk script using system COMMAND in Operating System - HP-UX

awk script using system COMMAND

Glenn Morton_1 — Tue, 08 Oct 2002 18:42:04 GMT

I have an awk script that was reccommened on 10/7/02. See the my previous entry, for details.
Here is the line that I'm having a problem with.

COMMAND=("grep -q " substr($0,53,4) " /tmp/reffile")

The above line works, but it needs to have a variable for the file name, based on SITECODE. I tried $REFFILE, but it didn't work. This script will use different DATAFILE files, and REFFILE files, based on SITECODE.
See the attachment for the exact coding.
The attachment has additional comments.

Re: awk script using system COMMAND

Sridhar Bhaskarla — Tue, 08 Oct 2002 18:54:15 GMT

Glenn,

If you checked that thread, I posted one script purely in Shell that can be easily modified to fit your purposes. It will take care of things like this.

-Sri

Re: awk script using system COMMAND

James R. Ferguson — Tue, 08 Oct 2002 19:18:36 GMT

Hi Glenn:

You can pass a variable into a 'awk' script thusly:

#!/usr/bin/sh
typeset THEFILE=$1
awk -v THEFILE="$THEFILE"'
{ if ($2~/455/) {
if (substr($4,17,6)=="abcdef") {
COMMAND=("grep -q " substr($4,1,4) " "THEFILE)
if (system(COMMAND))
{print $0 "XXbb"
}
else
{print $0 "XXYY"
}
close(COMMAND)
}
else
{print $0"bbbb"
}
}
else
{print $0
}
} ' /tmp/myinput

Regards!

...JRF...

Re: awk script using system COMMAND

wyan lowe — Tue, 08 Oct 2002 20:13:55 GMT

You might try this:

#!/bin/sh
#

FILE=$1
VAR=$2

echo "debug: file=$FILE var=$VAR"

echo "" | awk '{
printf "this is the '"${FILE}"' file.\n"
printf "and this the '"${VAR}"' variable\n"
}'

=======================
> ./pass_vars BLAH BLEH
debug: file=BLAH var=BLEH
this is the BLAH file.
and this the BLEH variable

> ./pass_vars "BLAH BLAH" "BLEH BLEH"
debug: file=BLAH BLAH var=BLEH BLEH
this is the BLAH BLAH file.
and this the BLEH BLEH variable

Re: awk script using system COMMAND

James R. Ferguson — Tue, 08 Oct 2002 20:29:21 GMT

Hi (again) Glenn:

...sorry, priority interrupt. In the case you have shown, I'd simply create a procedure (function) to which you pass the value of the input file (INPFILE) and the reference file (REFFILE). This will allow you to call the 'awk' code with variable arguments. Mind you, this is but one way you can implement this.

#!/usr/bin/sh
typeset INPFILE=$1
typeset REFFILE=$2
#
function filter
{
awk -v REFFILE=$REFFILE '
{ if ($2~/455/) {
if (substr($4,17,6)=="abcdef") {
COMMAND=("grep -q " substr($4,1,4) " "REFFILE)
if (system(COMMAND))
{print $0 "XXbb"
}
else
{print $0 "XXYY"
}
close(COMMAND)
}
else
{print $0"bbbb"
}
}
else
{print $0
}
} ' $INPFILE
} #_endof_filter
#
filter $INPFILE $REFFILE
exit 0

...since the indentation will be mangled here, I've also attached the above script.

Regards!

...JRF...