A single regex

kiat has asked for the wisdom of the Perl Monks concerning the following question:

Hi monks,

I'm wondering if it's possible to have a single regex that accepts a string if it's one of the following:

a) letters only
b) letters plus underscore
c) letters with numbers
d) letters plus underscore plus numbers

Strings that contain purely numbers or numbers with underscores are unacceptable.

Thanks in anticipation :)

Update: Thanks! Funny, I kept thinking it must be done with two regexes. Btw, from Re: A single regex, I'm wondering if it would still be doable with a single regex with one more constraint added:

Maximally one underscore.

Update: Wow, regex is really quite an art and a science in itself. Thanks to all for helping :)

Comment on A single regex

Replies are listed 'Best First'.
Re: A single regex by ccn (Vicar) on Sep 10, 2004 at 14:29 UTC
`/^\w[A-Za-z]\w$/` It means at least one letter and any optional numbers or undescores	[reply] [d/l]
Re: A single regex by BrowserUk (Patriarch) on Sep 10, 2004 at 14:34 UTC
Add some more testcases to this: `m[^(?=.*[a-zA-Z])[\w]+$] and print "$_ : ok" for qw[ 123 123_ 123_a a_ a_123 abc _ a123]; 123_a : ok a_ : ok a_123 : ok abc : ok a123 : ok` [download] Tweaked: The \d was redundant. Examine what is said, not who speaks. "Efficiency is intelligent laziness." -David Dunham "Think for yourself!" - Abigail "Memory, processor, disk in that order on the hardware side. Algorithm, algorithm, algorithm on the code side." - tachyon	[reply] [d/l]
Re: A single regex by Roy Johnson (Monsignor) on Sep 10, 2004 at 15:27 UTC
With the maximum-of-one-underscore constraint: `/^[a-z0-9](?:[a-z][a-z0-9]_?\|_[a-z0-9][a-z])[a-z0-9]$/i` [download] It's either going to have a letter before an (optional) underscore, or it's going to have an underscore before a letter. Surrounding and intervening characters can be letters or numbers. A lookahead solution would check that there's no more than one optional underscore, and that there's at least one letter: `/^(?=[^_]_?[^_]$)\w[a-z]\w$/i;` [download] In both cases, you're kind of doing multiple regexes. Caution: Contents may have been coded under pressure.	[reply] [d/l] [select]
Re: A single regex by Sidhekin (Priest) on Sep 10, 2004 at 15:49 UTC
"Maximally one underscore", like "maximally n anything", is easiest done adding a negative lookahead after the start anchor: `/^(?!._._)\w[A-Za-z]\w$/` [download] (That is also the most readable way, IMO, and not too slow, though if execution speed is an issue, doing it as a single regex is probably wrong anyway.) `print "Just another Perl ${\(trickster and hacker)},"` The Sidhekin proves Sidhe did it!	[reply] [d/l] [select]
Re: A single regex by BrowserUk (Patriarch) on Sep 10, 2004 at 15:22 UTC
I'm wondering if it would still be doable with a single regex with one more constraint added: Maximally one underscore. Yes. `m[^(?!=._._)(?=.*[a-zA-Z])[\w]+$]` [download] Corrected. Thanks Roy Johnson Examine what is said, not who speaks. "Efficiency is intelligent laziness." -David Dunham "Think for yourself!" - Abigail "Memory, processor, disk in that order on the hardware side. Algorithm, algorithm, algorithm on the code side." - tachyon	[reply] [d/l]
Re: A single regex by graff (Chancellor) on Sep 10, 2004 at 17:14 UTC
With the limit of zero or one underscore, it becomes a case where I'd prefer two separate conditions -- it just seems easier (and is probably faster): `( tr/_/_/ <= 1 and /^\w[A-Za-z]\w$/ )` [download]	[reply] [d/l]
Re: A single regex by Anonymous Monk on Sep 10, 2004 at 14:42 UTC
`/^[[:word:]][[:alpha:]][[:word:]]$/` [download] This accepts words with two underscores, it isn't clear from your description whether more than one underscore is allowed. Basically, your requirements reduce to: letters, numbers, underscores, with at least one letter. Which means you have zero or more letters, numbers, and underscores, then a letter, then again zero or more letters, numbers, and underscores.	[reply] [d/l]
Re: A single regex by Crian (Curate) on Sep 10, 2004 at 14:31 UTC
someone else was faster	[reply]