[Solved] StringRegExp, weird case.

[MvGulik]

whatever Edited February 7, 2011 by MvGulik

[MvGulik]

whatever Edited February 7, 2011 by MvGulik

[Richard Robertson]

I'm afraid I'm having trouble following what's going on here. What exactly is '[0:\0][1:\1][2:\2][3:\3][4:\4][5:\5][6:\6][7:\7][8:\8][9:\9]'?

[MvGulik]

whatever Edited February 7, 2011 by MvGulik

[Bowmore]

This produces the example output you gave. However I don't really understand what the you are trying to achieve in general so it may not work in with more generic examples or where you have more than 9 groups on a line. More examples of input and expected output would be useful.

Func MAIN()
    Local $sPattern = '([^,]*),([^,]*),?([^,]*),?([^,]*),?([^,]*),?([^,]*),?([^,]*),?([^,]*),?([^,]*)'
    Local $sSource = 'word1,word2,word3'
    ConsoleWrite('1: = ' & RE_Debug($sSource, $sPattern) & @CRLF)
    ;; output: "[x:...]" = captured part, at level x.
    ;; 1: = [0:word1][0:],word2,word3
EndFunc

Func RE_Debug($sSource, $sPattern)

        $aSource = StringRegExp($sSource, $sPattern,3)
        $sSource = StringRegExpReplace($sSource, $sPattern, '[1:\1][0:],[2:\2][0:],[3:\3][0:],[4:\4][0:],[5:\5][0:],[6:\6][0:],[7:\7][0:],[8:\8][0:],[9:\9][0:]')
    If @error Then Return SetError(@error, @extended, '')
    $sSource = StringRegExpReplace($sSource, ',\[[1-9]:\]\[[0]:\]', '')
    If @error Then Return SetError(@error, @extended, '')
    Return $sSource
EndFunc
MAIN()

Edited February 7, 2010 by Bowmore

[MvGulik]

whatever Edited February 7, 2011 by MvGulik

[Bowmore]

Does changing you match pattern to ''[^,]+'' help?

[MvGulik]

whatever Edited February 7, 2011 by MvGulik

[Richard Robertson]

What's this line for? It seems to be replacing the text with the text found, basically doing nothing.

$sSource = StringRegExpReplace($sSource, '\[[1-9]:\]', '\1')

As for why you aren't getting a 0 width match, it's probably because AutoIt is removing it. Very rarely is a 0 width match useful.

Edited February 7, 2010 by Richard Robertson

[jchd]

I fully agree with MvGulik that the [^,]* pattern definitely sounds like if it should match all three word* instance, and it does with another PCRE engine.

For instance:

C:\Documents and Settings\Jean-Christophe>pcretest-7.8
PCRE version 7.8 2008-09-05

 re> /[^,]*/g
data> word1,word2,word3
 0: word1
 0:
 0: word2
 0:
 0: word3
 0:
data>
 re>

The pattern being that simple, is it possible it's due to a PCRE build option combination used for making this module in the AutoIt build?

Still thinking, it escapes me!

Edited February 8, 2010 by jchd

[MvGulik]

whatever Edited February 7, 2011 by MvGulik

[jchd]

The closest I can get to what you want (at least as I understand it!) is this way:

Local $sPattern = '([^,]*)(?:,?)'
    Local $sSource = 'word1,word2,,word3,,'
    _ArrayDisplay(StringRegExp($sSource, $sPattern, 3))
    ConsoleWrite(StringRegExpReplace($sSource, $sPattern, '[0:$1]') & @CRLF)

Does that fit your bill?

Edit: sorry, no it doesn't really fit, as it will create a dummy empty capture at the end if the source doesn't have a trailing comma.

You can still easily circumvent this by appending a comma to the source anyway, and remove or ignore that last empty capture.

That's always a problem when you have intermediary separators, e.g. CSV That you create it or parse it, you always have to treat the first (resp. last) differently, due to no leading (resp. trailing) seperator. Regexps don't bent into this N+1 vs. N pattern naturally.

Edited February 7, 2010 by jchd

[Malkey]

....

I have this string: "word1,word2,,word3",

and like to get the following output: "[0:word1],[0:word2],[0:],[0:word3]"

....

This is getting off topic, but if we are playing magic parlor tricks, then this works.

Local $sSource = 'word1,word2,,word3'
Local $sPattern = '(?:[^,]+)|(?:,{2})'

ConsoleWrite(StringReplace(StringRegExpReplace($sSource, $sPattern, '[0:$0]'), "[0:,,]", ",[0:],") & @CRLF)

[MvGulik]

whatever Edited February 7, 2011 by MvGulik

[Richard Robertson]

Just to be clear, you're trying to match any text between commas, yes? And then display each one using a "[0:match],[0:next match]..." style?

Why not get each match separately and then concatenate the results? Wouldn't that be a much cleaner and easier to understand solution?

[MvGulik]

whatever Edited February 7, 2011 by MvGulik

[MvGulik]

whatever Edited February 7, 2011 by MvGulik

[Richard Robertson]

And then display each one using a "[0:match],[0:next match]..." style?

Hum, No. Thats just the RE_Debug() support part.

Why not get each match separately and then concatenate the results?

I don't think I'm following you here. But it sound slower than what I already have. ←#16

I didn't realize that you were using it to preview data. I thought that you were trying to format it that way from the beginning. My bad.

[MvGulik]

whatever Edited February 7, 2011 by MvGulik

[jchd]

If you can get along with ignoring empty fields altogether, then this is simpler of course.

In most of the use I have of this kind of stuff I can't do that since it would shift the data in the wrong database columns.

For example if the incomming data is:

123,"text field ""one""", ,"",2009/12/17,,"another text"

then I need to get exactly (I prefer to deal with all in text format):

'123'

'text field "one"'

'<null>'

''

'2009/12/17'

'<null>'

'another text'

I've yet to find a reasonable way to do it in one shot (and it would break anyway in case some dummy is creative enough to insert comments as he sees fit).

123,;you can see a nice comment here;"text field ""one""", ,"",2009/12/17,### yet another "comment" <style> I have encountered!! # ,"another text"

If you look around, you find there are so many variants of these "creative" formats that no definitive solution exists.

If only people would stick to RFCs and published reasonable specs!

For comma separated values:

http://en.wikipedia.org/wiki/Comma-separated_values

http://tools.ietf.org/html/rfc4180

For tab separated values: http://www.fileformat.info/info/mimetype/text/tab-separated-values/index.htm

http://www.cs.tut.fi/~jkorpela/TSV.html

Cheers,