[cdelgadop]

Hi

I have a LPAR running shell scripts using sed command to make some convertions in datafiles. When those programs run, performance at CPU level goes down almost at 100% user. I'm not such a expert in sed so i need your help in order to improve those programs. Following one sample of those sed sentences:

cat encabezado.tx $1 |sed "s/\| /\|/g" | sed "s/,/./g" > $1_ok

The program above looks for every blank space after the pipe character and change it by no space at all, then changes comma character by dot. But i think this can be improved. I think instead of having sed to change the blank spaces and then replacing commas, all these tasks can be exec at the same time.

Any help please.

[Mark Taylor]

My sed is crud also, I prefer to use perl .. and, this can also probably be improved into one substitution instead of seperating 2 subs with a ";" because my perl is not that good either, but htis works ok.

perl -pe "s/\| /\|/g;s/,/./g" your_text_file_name.txt

HTH
Mark Taylor