I have a problem like this:

Suppose I have this text: (the part between --------)

--------------------------------------------------------------
1 ENTRY --- one white space between "1" and "ENTRY"
2 TEST --- one white space between "2" and "TEST"
3 ENTRY --- two white spaces between "3" and "ENTRY"
4 TEST --- two white spaces between "4" and "TEST"
--------------------------------------------------------------

Perl style. Regular expression

[0-9][ \t]+[^E]

matched line 2, 3 and 4 but not line 1. I expect it to match only 2 and 4.

Why line 3 got matched?

For line 3, seems that ultraedit perl used [ \t]+ to match one white space, and used [^E] to match the 2nd white space.

Anyone know how to generate a perl style regular expression to match only line 2 and 4 but not line 1 and 3? That is, do not match lines with one or more blanks following the number [0-9], followed by "ENTRY"?


Thanks.

You're right, the original regex matches line 3 because [^E] matches the space character. Greedy does not mean that a regex part won't give up a (partial) match if the overall regex requires it.

A more robust way to circumvent this would be to use as your regex. Now the space/tab bit is enclosed in a so-called atomic group that will not be backtracked into once it has matched successfully. In other words, it will "use up" the space/tab part of the match and not try to go back into this if the following character is not an E.