Shell Programming and Scripting

I'm working with a large file with multiple records, each record begins with ISA[^a-zA-Z0-9]. The issue is, sometimes ISA[^a-zA-Z0-9] is at the start of the line, sometimes it's in the middle of the line. So before I can csplit my main file into multiple records, I have to get each record header onto its own line.

I've been having trouble using sed or tr to do this. Do I have to capture each occurence of ISA[^a-zA-Z0-9] into a variable and then use the variable as part of the regex in a sed statement?

I'm open to all suggestions on how to accomplish this. This is my first time posting, so I apologize if I'm not asking in the forum style or detail.

Thanks!

Note: By ISA[^a-zA-Z0-9] , I'm meaning ISA(+ any non-alphanumeric character) ... in case there's a better way.

Try:

Code:

perl -p0e 's/ISA[\W_]/\n$&/g' file

bartus11 ... that worked beautifully, thank you so much ... now I'm rethinking my whole script!

Is there a \W equivalent for non-printable or non-ascii characters?

As far as I know \W matches all non-alphanumeric characters, including non-ASCII and non-printable ones.

Thanks bartus11, this is an example of a record ...

Code:

ISA~00~ ~00~ ~ZZ~TPM ~ZZ~TPU~110301~0037~U~00200~000002452~0~P~<GS~FA~TPY~TPM~110301~0037~2470~X~004010ST~997~0932AK1~SQ~110456465AK2~866~0001AK5~AAK2~866~0002AK5~AAK9~A~2~2~2SE~8~0932GE~1~2470IEA~1~000002452

The main file has many of these records. The ~ in this case can be any non-alphanumeric value and that's fine, they can stay. In this case  (Hex 1C) needs to be replaced with newline. Throughout the file, the IC can be any non-ascii, non-printable character.

My original thought was to put the value at position 63 into a variable and replace it with a newline with sed but from your code example, I can see there are much better and concise ways to handle substitutions.

I haven't had a chance to experiment with perl and [^[:ascii:]] or [^[rint:]] ... will either work? Is there an entirely different way?

Thanks again for taking time

So you want to replace all the "hex 1C" characters with newline in your file? Or any non-printable character?

any non-printable character with a newline... IC was just an example