That’s a good point about context. I look at it as having one more tool in the belt. Linux/Unix have so many specialized tools at the user’s disposal (as well as more generalized tools like perl, python, ruby, c, et..) that it is a shame to specialize in only one.

The title is the intentionally-inflammatory “Throw out your Perl,” so is it really surprising that someone felt the need to respond?

To someone familiar with both languages, they are both about equivalently readable. Perl’s “$” are ugly, but then again, awk uses “$” for the column contents, so overall the one-liners seem to be about equally obfuscated to the casual viewer.

http://ipython.scipy.org/moin/Cookbook/StringListProcessing

So far, fields() does not support arbitrary separator (whitespace only) but than can be fixed

The “better” perl version for stocks doesn’t aggregate the profit/loss by stock:

ABC 698
FOO -656
BAR 271.5
BAZ -472.5
NI 234.6
ABC 1200
Net: 1275.6

Then again, the awk version gets it totally wrong for ABC:

FOO -656
ABC 1200
BAZ -472.5
NI 234.6
BAR 271.5
Net: 1275.6

Seriously, those are nice one-liners. I should’ve noted that I also do know, use, and teach perl (only upon request, by gunpoint — but it happens). The article makes it pretty clear, though, that I’m talking about awk in a particular context, for a given purpose. You Perl hackers should know all about context. It’s not a blanket “awk beats perl” thing – I just think awk is cleaner and easier for these types of tasks. Those people I’ve been able to show awk solutions to, who were using perl, now don’t use perl for those tasks. Not because you can’t do it in perl (duh), but because you don’t have to, and awk is a bit more intuitive for certain tasks.

I have plenty of other posts I have yet to write about why I hate perl, and why I think it will eventually become the next cobol — save your flames for those future posts.