Regex: select all possible of a pattern

[dlite922]

I'm putting this question in perl forum be we don't have a regex forum (we should) and 2, you guys rock at Regex.

I'm trying to find all patterns of a number to accommodate a dyslexic typer. By matching any pattern that contains the characters in the given input as many times as they /exist/ in the input. English sucks, example's better:

given 231

match 2 3 and 1 in any order, but do not repeat because there's only a single 1, a single 2, and a single 3:

123
132
213
231
312
321

but not:
111
112
221
222
311
333
223
332
etc...

My brain is locked. I'm thinking it should be easy but all i can come up with is along the lines of:

 (3|2|1)+ 

which of course matches ALL of the above unwanted outputs.


Thanks,



Dan

[nithinpes]

You can try this:

/^(\d)(?!\1)(\d)(?!\1)(?!\2)\d$/
 

The negative lookahead assertion(?!) checks that the character that follows does not match the pattern mentioned inside the cluster.

[dlite922]

Thanks, ! suck at neg lookaheads.

I did change the \d to [123] though, because my character set must be those. Otherwise 456 would get through also.

^([123])(?!\1)([123])(?!\1)(?!\2)[123]$

Dan