Cant figure out something with StringRegExp

[sugi]

Hello,

I've got a problem with a regular expression. I have a two lines: "foobar skill" and "something else skill". Now i want to extract this "foobar" or "something else".

So i created this regular expression: "(\w+( \w+)?) skill"

In my reading of this RE this means:

match the first word

maybe have a space followed by another word

a space

the word "skill"

According to the regex coach this expression should work, but if the text has three words in total it does not work in AutoIt as the following code shows:

$text="foobar skill"
$regexp='(\w+( \w+)?) skill'
$x = Number(StringRegExp($text, $regexp, 0))
MsgBox(64, 'RegExp?', $x)

$text="something else skill"
$regexp='(\w+( \w+)?) skill'
$x = Number(StringRegExp($text, $regexp, 0))
MsgBox(64, 'RegExp?', $x)

But if I change the "?" to a "*" it works. Which is strange I think. After all the "?" should mean "zero or one" (and in case of the two word line it should mean zero) and the "*" means "zero or more" (and in case of the two word line it means zero). So both should be give the same result for the lines I used to test them.

I cant figure out what's wrong in my expression. Do you have any idea?

[LxP]

I'm not entirely sure of what you're trying to do but perhaps this will help:

StringRegExp('something else skill', '(.+?)\w+skill')

(Match anything that occurs before some whitespace and the word 'skill'.)

[Francis Lennert (France)]

Hello All,

Maybe that will help. If you want to extract from strings, I think that you must take a look

at the last Flag (3) from the StringRegExp command. That will create an array with values

which are simple to use. The ( and ) will help you to create Blocks which will be values in the Array.

HTH,

Francis

Dim $i
Dim $Text[2]
Dim $Tokens 

$text[0]="foobar skill"
$text[1]="something else skill"
$regexp="^(.*)( skill)"

for $i=0 to 1
    $tokens = StringRegExp($text[$i], $regexp, 3)
    MsgBox(64, 'RegExp?', $tokens[0])
next

[sugi]

I know about the StringRegExp flag and that 0 only tells you if there's a match or not. I'M only using 0 because I wanted to validate if it matches at all because I had strange results.

Simply using ".*" doesnt help me as the actual line is a lot longer. I shortened it to the actual problem I have. so neither ".*" or ".+?" is usable for me. I need to match "a word", possibly followed by "a space and a second word" and always followed by " skill".

[LxP]

How about this then:

StringRegExp('something else skill', '(\w+\s+){1,2}skill')

[sugi]

StringRegExp('something else skill', '(\w+\s+){1,2}skill')

That works if I add a StringStripWS as I dont want the last space. But I'm still trying to figure out why my regex doesnt work. After all at least one program accepts them, so it cannot only be a wrong regex.

[sugi]

Funny thing: I just tried this code:

$text="Someone's fishing skill rises 0.2 points.1"
$regexp='\w+.s ((\w+\s+){1,2})skill rises (\d+.\d+) points\.1'
$x = Number(StringRegExp($text, $regexp, 0))
MsgBox(64, 'RegExp?', $x)

It tells me it doesnt match. But if I use

$regexp='\w+.s (\w+\s+){1,2}skill rises (\d+.\d+) points\.1'

It works. It looks like AutoIt does not support nested groups.

[Francis Lennert (France)]

Dim $i
Dim $Text[4]
Dim $Tokens 

$text[0]="foobar skill";  Good
$text[1]="something super else skill"; not good
$text[2]="something else skill";  good
$text[3]="something else  skill"; good 


$regexp="^(\w+\s{1,2}){1,2}skill"

for $i=0 to 3
    $tokens = StringRegExp($text[$i], $regexp, 0)
    MsgBox(64, 'RegExp?', $tokens)
next

???????????????????

seems to work

[sugi]

$regexp="^(\w+\s{1,2}){1,2}skill"

That regex has a different meaning: A word, followed by one or two spaces. Your text[3] should NOT work.

I'm talking about groups within other groups because I dont want to use postprocessing at all when the regex is able to give me what i want.

Edited November 26, 2005 by sugi