CODEWITHSUNDEEP
regexunixyacclex
footy
footy

Reputation: 5911

simple regular expression question

How to match aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaab where number of a's should be min of 10?

I mean i know this way:

[a][a][a][a][a][a][a][a][a][a][a][a][a]a*b

But there must be a better elegant method where is if my min number of a's become say 100..

What is it? I am trying to match (a^n)b sort of thing where n can be anything

EDIT:

I forgot to mention this is done using lex and yacc.. where the lex has to return a token to yacc.

%{
#include "y.tab.h"
%}
%%
aaaaaaaaaa[a]*b {return ok;}
\n {return '\n';}
. {return 0;}
%%

Upvotes: 0

Views: 141

Answers (4)

Ise Wisteria
Ise Wisteria

Reputation: 11669

If your lex is flex, you can use a{10,}.

If not so, according to 3. Lex Regular Expressions , you can use a{10}a* instead.

Upvotes: 2

corlettk
corlettk

Reputation: 13574

Footy,

[WARNING: This answer is COMPLETE BUNKUM!!!]

(if you mean soccer, we're swarn enemies ;-)

Ummm, No... That is not as far as I know, using "the standard" regular expression syntax as supported by sed, grep, nawk, and the likes... and no not even egrep... As far as I know, the a{10,*} syntax (which is exactly what you're hankering for) didn't emerge until Perl rewrote all the books on the capabilities of regular expressions... and (don't quote me on this) I don't think that happened until like version 5.

So yeah, If you're stuck with using nawk, then it's the aaaaaaaaardvarking hardway dude. Sorry.

Cheers. Keith.

EDIT:

Hmmm... I seem to be the odd-man-out here... maybe everone-elses "standard operating environment(s)" have been updated with "standard tools" that recognise later regular expression syntax extensions... Sooo... Hmmm... I tested this on my (three year old) cygwin implementation of egrep... and it suprised me by actually working!!!

Administrator@snadbox3 ~
$ egrep 'a{3,}b' <<-eof
> ab
> aab
> aaab
> aaaab
> eof
aaab
aaaab

So I'm WRONG all ends up... looks like the "new" {min,[max]} syntax is reasonably well supported, and I'm getting old. Sigh.

Cheers. Keith.

Upvotes: 1

Ali1S232
Ali1S232

Reputation: 3411

use this format : a^na*b and replace n with any number you want.

Upvotes: -1

aioobe
aioobe

Reputation: 420971

Try

a{10,}

which says a 10 or more times.

grep -E "a{10,}" filename

matches aaaaaaaaaaaaaaaaaaaaaaaaab but not aaaaaaaaab.

Upvotes: 7

Related Questions