Shell Programming and Scripting

AIX version 5.2 has an arbitrary limit of 399 on the length of regular expressions and literal strings in an Awk program.

Apart from splitting down the string into multiple strings is there any way around this annoyance?

I have a search string that ends up way way too long and blows out awk!
awk: 0602-508 The string or regular expression beginning with
the characters "^.*[.]DLC$" cannot contain more than 399 bytes. The source line is 1.
The error context is
>>> .]DEL$|^.*[.]ONU[.]DEL$|^.*[.]OSU[.]DEL$|^.*[.]ORJ[.]DEL$|^.*[.]RNJ[.]DEL$|^.*[.]RNU[.]DEL$|^.*[.]RNA[.]DEL$| <<< ^.*[.]RRJ[.]DEL$|^.*[.]ONJ[.]DEL$|^.*[.]OBT[.]DEL$|^.*[.]AWH[.]DEL$|^.*[.]AWI[.]DEL$|^.*[.]MAJ[.]DEL$|^.*[.]MAV[.]DEL$|^.*[.]MDA[.]DEL$|^.*[.]MDC[.]DEL$|^.*[.]MDN[.]DEL$|^.*[.]MFC[.]DEL$|^.*[.]MFN[.]DEL$|^.*[.]MFV[.]DEL$|^.*[.]RSU[.]DEL$|^.*ERR.*$|^.*[.]DRD[.]DEL$/){print $NF}}
Syntax Error The source line is 1.
awk: 0602-502 The statement cannot be correctly parsed. The source line is 1.

Any advice would be much appreciated.

I know of no way to come around this limitation - BUT:

do you really think that what you have pasted constitutes professionally crafted code? Honestly, ask yourself: if i look in one year from now onto this code, would i be able to know what i have done? Now suppose someone else is maintaining this code, he would not even have the advantage of being you and knowing how you usually do things - he will be lost completely.

Do yourself (and any possible maintainer) a favor and rewrite it. Take the awk limitation not as a nuisance but as a chance.

Copy your script here and some example input file and desired output and we can help you write better code.

I hope this helps.

bakunin

PS: coming to think about it please post it to the "shell scripting and programming" part of the forum because this would be not in the scope of the AIX board any more.

The string is built up by grabbing a list of search values from config files, that's certainly not hardcoded, I'd need to be shot if it was

FILEs="`get_configs | grep -v "^#" | grep ",$APP_NAME," | awk -F, {'print $3'} | tr '\n' '|' | sed 's/\./[.]/g; s/*/.*/g; s/?/./g; s/|$//; s/^/^/; s/|/$|\^/g' `$"

for FILE in `awk '{if ($NF~/'${FILES}'/){print $NF}}' $temp_dir_file`
do
.....
done

This builds up the list of values for the specific APP_NAME, unfortunately the one in question has reams of bloody entries for it!

something along these lines..

Code:

nawk -v pats='^.*[.]ONU[.]DEL$|^.*[.]OSU[.]DEL$' '
   BEGIN {
      n= split(pats, patsA, "|")
   }
   {
      for(i=1; i<=n; i++)
         if ($0 ~ patsA[i]) {
            print $NF
            next
         }
   }
' myFile

You're right, sorry about that. Could a mod move the thread to the correct forum for me and rename it to "Awk limitations - 399 byte limit" so as to avoid cross posting please.