Non greedy search in bash
 

Is there a way to search non greedily as we do in perl??
Say I have a string
str="Abcdefgcdhd";
Now I want to replace "defgcd" with "rgt"
In perl I can use

str =~ s/d.*?d/rgt

But with bash how can I acheive this non greedy search.Is there any .*? in bash??



Thanks
Rajesh KSV

i'm not entirely sure what you mean.

you could try sed


str='abcdefgcdhd'
echo $str|sed -e 's/defgcd/rgt/'


hope this helps

Hop over to http://tldp.org and get the Bash Guide for Beginners. Later, get the Advanced Bash Scripting Guide.

At some point, Regexes are all the same---but they are used differently in different utilities. And of course there are "standard" and "extended" flavors

What is the precise meaning of "*?" in PERL?

non greedy means: for example
you want to match words inside <>. Not from the first < to the last >.

bash doesn't yet support non greediness, so is sed /awk (correct me if i am wrong).

the other way is to do pure string manipulation.

@OP, show examples of input and output you want..

Thanks to all for your quick replies.
I have an html page which contains a table

<table><tr><td>cell1</td><td>cell2</td></tr>
<tr><td>cell3</td><td>cell4</td></tr>
</table>

I want to delete the first row.I dont know what will be contents of cells.

sed 's|<table><tr>.*</tr>|<table>|'

doesnt work.What should I do?? ----I.e,how to achieve Non greedy search

And one more thing I want to search for words which contain "ba" but not followed by "la" .i.e., I want to search for a specific pattern which shouldnt be followed by another specific pattern.How can I write a regex for that in awk/sed

Thanks in Advance

not sure if the output is what you want though. the method is split by "</tr>" as the record separator since this denotes end of one row.

write this in awk

This is a multi-line editing task and none of the options you list are suited for it (all of them can be made to do multi-line edits, but only by selling your soul to some mythical creature). It's time to learn how to write your scripts in something other than Bash. Options:

Ruby
Python
Perl

Wait a minute. If you already know how to write in Perl, why are you trying to perform this obvious Perlish task in Bash? This is what Perl was intended to do -- its purpose in life. The same could be said of Python and Ruby, with the difference that the latter languages are easier to read.

@lutsup ; Thanks for your suggestion.But are there tools in perl which are as good as grep/awk ??
Any good links/books for learning perl quickly :)

Perl is at least as good, if not better. In fact, many tools use the PCRE - Perl compatible regex engine.

http://perldoc.perl.org/
http://www.perlmonks.org/?node=Tutorials

Perl Cookbook: highly recommended.

Good links.That perl cookbook is really good.But still I find difficulty in using perl.How can I grep through gz files in perl (without compressing them)

In bash I can say


count=$(zgrep "Error" /logs/2009/10/12/*/*gz | wc -l)


But in perl its looks like a big job.First I have to read each file using Archive::Zip and then use grep on each file.I searched on net but was not able to find a better solution.Can I use system() and use the above command :) I know that if I do that I am just using bash with perl as a beautiful wrapper .Am I missing anything here??


Thanks in Advance

Try any of
http://search.cpan.org/~pmqs/IO-Comp.../AnyInflate.pm
http://search.cpan.org/~softdia/Tie-...ib/Tie/Gzip.pm
http://search.cpan.org/~nwclark/Perl...p-0.17/gzip.pm

I think prob 1 or 3, but you could just go to search.cpan.org and enter 'gzip' in the search box.
Anyway, that's what I did ;)

yes of course, but why do that if you want to use Perl? you make your code non portable.

yes you are. like what i mentioned above. However, for text parsing like in your case, awk is as good as Perl and sometimes better so if you are not comfortable with Perl yet, stick back to what you know. Many things can still be done easily with the shell.

Say what? Perl is 100 times better for text processing than grep times akw raised to the third power. No contest. Grep and awk are weak imitations.

Google is your friend.

Yup I can go to awk but I find things in my project which I cant do with awk easily as I can do with perl.. I was actually searching for "search.cpan.org" sort of site only.Thanks crism :)

I must take issue with this characterization.

First off, perl is certainly great. It brings a lot of overhead, though . . . for writing simple shell scripts or constructing effective one-time commands, grep, sed, and awk are fan-friggin-tastic. They are, however, certainly not to be wasted on the profane.

When used in concert, grep, sed, awk can do almost anything with text.

They also all three pre-date perl, which shows up for the first time in 1987, and so it is impossible for them to be imitations, weak or otherwise.