RegExReplace glitch... Topic is solved

[kunkel321]

Hi Folks,
My regex here seems to work some of the time when testing this small block of code. It seems to always fail in my larger code though...

Code: Select all

MyEntry := "They uses the following. They reads often."
MyEntry := RegExReplace(MyEntry, "([Tt]hey\s)(.*)(s)", "$1$2")
MsgBox post fix:  %MyEntry%

It's supposed to do a simple grammar fix: Change uses > use and reads > read. Oddly, when I test it, it usually changes reads but not uses. Any ideas what I've done wrong?

EDIT: Note also, that this one DOES work

Code: Select all

MsgBox % RegExReplace("They uses the book", "([Tt]hey\s)(.*)(s)", "$1$2")

This would seem to be functionally identical -- yes?

EDIT2: Nope... If I swap the haystack strings, then the error follows the longer string... So I guess my regex likes the shorter one, but not the longer one.

[mikeyww]

The regex "expands" by default. It is called "greedy"-- eating up as many characters as possible, while maintaining a match. The following would work.

Code: Select all

MyEntry := RegExReplace(MyEntry, "i)they \w+\Ks\b")

Or the "lazy" approach, which finds the "shortest" match:

Code: Select all

MyEntry := RegExReplace(MyEntry, "([Tt]hey\s)(.*?)s\b", "$1$2")

The "?" after the wildcard (*) tells the parser to be lazy.

[kunkel321]

Ah yes, of course... It was matching everything between the first "They" and the last word ending in "s." Thanks Mike!