Regex captures one instance of the pattern but not the other. How to fix this?

Question

I have this piece of text:

W/NNP Yes/NNP Get/NNP Paid/NNP for/IN Going/NNP to/TO College/NNP !/. Check/NNP it/PRP out/RP here/RB !/. http/NN :/: //sldollar.notlong.com/JJ apple/NN iphone/NN TGIF/NNP swine/NN flu/NN

I am currently using this regex to capture some regions of interest:

[a-zA-Z]*/NN[PS]* [a-zA-Z]*/NN[PS]*

I am using RegexPal to test this.

enter image description here

This captures TGIF/NNP swine/NN but not swine/NN flu/NN. Any suggestions on how to fix my regex to capture this?

@Maulwurfn: Yeap. Thank you. Just realized this can be solved using positive lookahead and posted my answer. — Legend, Aug 12 '12 at 06:23
http://stackoverflow.com/questions/3027718/overlapping-matches-with-finditer-in-python — , Aug 12 '12 at 06:23
@tripleee: Well something like that, just that I'm not tacking spam though ;) — Legend, Aug 12 '12 at 06:50

score 1 · Answer 1 · answered Aug 12 '12 at 06:22

1

In case anyone else needs this, I guess the answer is to use a positive lookahead:

([a-zA-Z]*/NN[PS]* )(?=([a-zA-Z]*/NN[PS]*))

answered Aug 12 '12 at 06:22

Legend

113,822
119
272
400

score 1 · Answer 2 · answered Aug 12 '12 at 06:22

Multiple matches cannot overlap.

apple/NN iphone/NN TGIF/NNP swine/NN flu/NN
AAAAAAAAAAAAAAAAAA BBBBBBBBBBBBBBBBB
         CCCCCCCCCCCCCCCCCC DDDDDDDDDDDDDDD

Matches marked A and B above follow each other, but because piece C starts in the middle of match A (and likewise for D within B), they are not matches.

You need to match once and then re-search at some point after the previous starting point, or use lookahead so the latter part isn't consumed.

Regex captures one instance of the pattern but not the other. How to fix this?

2 Answers2