Re: regular expressions in Prolog

On Mon, Sep 16, 2013 at 10:34 AM, John Benson <[email protected]>wrote:

>  However, when I used it I got the following results:
>
>     130 ?- amatchesregexb('ab12cd', '.*([0-9]+).*', Result).
>     Result = ['2'].
>
>     131 ?- amatchesregexb('ab12cd', '[^0-9]*([0-9]+).*', Result).
>     Result = ['12'].
>
>     132 ?-
>
>
> *I expected the first example capture **([0-9]+) to cap**ture '12', but
> it only captured '2'.*
>
> I had to add [^0-9] infront of the numeric capture to successfully
> capture both digits '12'.
>
> I'm a Prolog newbie, and don'tunderstand the code that I copied.
>

This regex behavior is correct and isn't specific to Prolog.  A * matches
greedily.  That is, it consumes as many characters as possible while still
allowing the whole pattern to match.  In the first example, .* can consume
the 1 and the entire regex still matches.  Using [^0-9]* prevents that part
of the pattern from matching the 1.

-- 
Michael
-------------- next part --------------
HTML attachment scrubbed and removed