awk script to remove invalid chars in a variable length record

[confuseddddd]

Used this script to remove the asterick from the 2nd field where sometimes contains an asterick.
> awk 'BEGIN { FS=OFS="*" }
> {
> if ($1=="CLM&quot {
> while ( $3 !~ /^[0-9]*[.][0-9][0-9]$/ ) {
> $2=$2 $3;
> for (x=3; x<split($0,a); x++)
> $x=$(x+1);
> }
> print $1,$2,$3,$4,$5,$6,$7,$8,$9,$10,$11,$12,$13,$14,$15,$16,$17,$18,$19,$20,$21,$22,$23,$24,$25;
> } else {
> print $0;
> }
> }' $(date -u +%Y%m%d).IN > $(date -u +%Y%m%d).001

However, because the fields in the record can vary, I am getting strange results:
original:
CLM*SANAR00 7072*65.00*WC**11^^1******EM^^^TX
after script
CLM*SANAR00 7072*65.00*WC**11^^1******EM^^^TX*************

original:
CLM*GALKA001 9778*295.00*WC**11^^1******EM^^^TX
after script:
CLM*GALKA001 9778*295.00*WC**11^^1******EM^^^TX*************

original:
CLM*TATMI004569*100.00*WC**11^^1******EM^^^TX********29
after script
CLM*TATMI004569*100.00*WC**11^^1******EM^^^TX********29*****

original:
CLM*37494-002*97.70*WC**11^^1******EM^^^FL
after script
CLM*37494-002*97.70*WC**11^^1******EM^^^FL*************

original:
CLM*001129S5 FWC*615.00*WC**11^^1******EM^^^FL
after script
CLM*001129S5 FWC*615.00*WC**11^^1******EM^^^FL*************

original:
CLM*350049582*35*388.00*WC**11^^1******EM^^^KS
after script
CLM*350049582*35*388.00*WC**11^^1******EM^^^KS*************

Anyone have any ideas on how to remove the unwanted asterick from the second field but still leave the rest of the record alone???

 

[PHV]

Try this:

}' $(date -u +%Y%m%d).IN |
 sed 's!\**$!! > $(date -u +%Y%m%d).001

Hope This Help
PH.

 

[Ygor]

Create an file called, say, &quot;rmstar&quot; containing ...

#!/usr/bin/sed -f
s/^\(CLM\*[^\*]*\)\*\([^\*]*\*[0-9]*\.[0-9][0.9]\)/\1\2/
s/^\(CLM\*[^\*]*\)\*\([^\*]*\*[0-9]*\.[0-9][0.9]\)/\1\2/
s/^\(CLM\*[^\*]*\)\*\([^\*]*\*[0-9]*\.[0-9][0.9]\)/\1\2/
s/^\(CLM\*[^\*]*\)\*\([^\*]*\*[0-9]*\.[0-9][0.9]\)/\1\2/
s/^\(CLM\*[^\*]*\)\*\([^\*]*\*[0-9]*\.[0-9][0.9]\)/\1\2/

Use chmod to make it executable, and use it like this....

rmstar $(date -u +%Y%m%d).IN > $(date -u +%Y%m%d).001

Tested....

< CLM*274*1979731*4.00*WC**21^^1******EM^^^IL
> CLM*2741979731*4.00*WC**21^^1******EM^^^IL

< CLM*350049582*35*388.00*WC**11^^1******EM^^^KS
> CLM*35004958235*388.00*WC**11^^1******EM^^^KS

 

[confuseddddd]

Okay, that worked great except it didn't catch all of the situations...

Have the following:
CLM*269605*73612*542.75*WC**11^^1******EM^^^NM
should be CLM*26960573612*542.75*WC**11^^1******EM^^^NM


CLM*30750****41.91*WC**11^^1******EM^^^TX
should be CLM*30750*41.91*WC**11^^1******EM^^^TX

CLM*210599*1*1*715.00*WC**22^^1******EM^^^TX
should be CLM*21059911*715.00*WC**22^^1******EM^^^TX

Any other ideas...I didn't see the s commands what was
different if anything. I thought I could just add another s command but wasn't sure....

 

[bi]

I'm not a sed or awk heavy user, but to those who are, regarding the original script in your post: is there some way to tell awk to &quot;print all the fields until you get to an end of line&quot; rather than have $1,$2,$3...$25? some of the lines don't have 25 fields and that seems to be why those extra * were being printed.

 

[vgersh99]

print $0

vlad
+----------------------------+
| #include<disclaimer.h> |
+----------------------------+

 

[Annihilannic]

bi,

If you needed to start part way through the line you could do something like this:

[tt]echo one two three four five | awk '{
for (i=3;i<=NF;i++) {
printf &quot;%s &quot;,$i
}
print &quot;&quot;
}'[/tt]


Annihilannic.