Regexp

DogmaticLefty

Anyone know how I can match the following words in my large text file:

aaaaaaaahhhhhhhhhhh
grrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr
ooooooooooooooooooccchhhhh
gggooooooooooooooggggggggggle
etc.

I'm trying to clean up some noisy text.

I've got something like s/[a-zA-Z]{3,}//g but it's not working...

If the same letter occurs more than three times, then the regexp should match so I can delete the word.

humbert

I think this tutorial is the best by a country mile and Expresso is a very handy tool for building up expressions.

http://www.codeproject.com/KB/dotnet/regextutorial.aspx

gggooogggle = g{3}o{3}g{3}le

g{1,3} matches between 1 and 3 repetitions. Dunno if that's exactly what you're looking for.

Oh and from experience I'd recommend always building up a regular expression gradually while running after each addition instead of writing it all and then wondering why you are getting no matches.

Oh, didn't fully read your post. You want a back reference like \S*(.)\1{2,}\S*

(.) matches any character and creates a numbered group labeled \1. Then \1{2,} looks for another two repetitions!

DogmaticLefty

Here is the answer:

$line =~ s/\b\w*(\w)\1{2,}\w*(?:\s|$)//g
Any words like aaaaaahhh or ggggggrrrrrr on the line are deleted.

I love line noise on vim terminals...