Shell Programming and Scripting

I am having to search for a specific string in multiple files. Problem is that the files contain wrapped data, so each basically has one huge record, and I get the grep error message of max line, etc.

I searched the forum, but didn't find a specific solution. I just need to command to spit out the file names that contain a specific string within its data.
I don't have perl or GNU. I am on AIX.

Help please!!!

So you can use:
grep -l "string" *.txt

If you are using EDI documents you can use the following to split the long string into individual lines based on segment terminators. Then on the resultant file you could perform your search for the pattern.

Code:

#! /usr/bin/ksh
FldDelim=$(head -1 $1 | cut -c4)
LnDelim=$(head -1 $1 | cut -c106)
cat $1 | tr "$FldDelim" '*' | tr "$LnDelim" \\n > $2

if the above script were named gerard1.sh you could call the same as
gerard1.sh in_file out_file

where out_file will contain line feed delimited file and then you could perform the grep on the out_file.

Jerardfjay

try fold command

to the person who suggested grep -l.........huh? Stil won't work.

OK, yes, I am working with EDI files, and I have Gentran software. So I decided to just write a small script to feed each file into a function, run Gentran's 'edifrmat' command that basically unwraps the data into a new file, searched the new file, and if it contained the string, copied the original file to where I need it.

Never heard of the 'fold' command, but thanks, I'll check it out!

fold command allows you to break a stream of characters based on the width parameter that you specify. The default is 80. However I dont think you can use it in this context.

jerardfjay is right. 'fold' is a general tools. for EDI files it has its own rules, using 'fold' may cause a word splited into two lines.