[MLUG] perl, awk, sed, etc.

[Mike Miller <EMAIL:PROTECTED>]

On Tue, 4 Dec 2007, Michael wrote:

> > I very frequently call perl from the bash script to do various little 
> > things, almost always regexp search/replace because that is what I 
> > know.
>
> Nothing wrong with that. Perl for short things and especially good for 
> regexp stuff. Of course awk and sed can sometimes be good for small 
> things too. :)

On SOLARIS, awk and sed can both be limited or buggy.  The GNU versions 
are clearly superior.  Example: awk can only count fields (NF) up to 
something like 256, which is not enough!  sed also fails without warning 
with long lines -- this bug is rather famous because it causes some 
installation scripts to fail on SOLARIS.

All I really know how to do with sed is search/replace, I just do it in 
perl instead.  I feel much more confident that it will work and the code 
will be portable to any system that has perl running.

I do use gawk a lot (but not awk, though I believe these are identical on 
a GNU/Linux system).  Mostly stuff like this:

gawk -F'\t' '$4>3 {print $2" "$5}' file.in > file.out

Mike

_______________________________________________
members mailing list
EMAIL:PROTECTED
http://mlug.missouri.edu/mailman/listinfo/members