RegExp to parse out text outside of ' or "

[jdelaney]

Hello all,

Trying to split a string into logical components.

I have string = "@testand='and or string' and .='and or string2' or contains(.='test or quote') and @testand='or string2'"

I want to split out the data by any 'or' that is not inside of a quote, or double quote.

Currently, the best/only way I know how, is to split by ALL 'or's, and then reconstruct based on counting ' or " to determin if inside or outside.

I know there is an easier way through regexp, hoping someone would help me out.

In the above string, i'm expecting an array with:

1) @testand='and or string' and .='and or string2'

2) contains(.='test or quote') and @testand='or string2'

Currently, the way I know how, is something like this:

expandcollapse popup

#include <Array.au3>
; split by all or's
$atest = StringSplit ( "@testand='and or string' and .='and or string2' or contains(.='test or quote') and @testand='or string2'", "or", 1 )
_ArrayDisplay ( $atest )
Dim $asHolder[1][2]
; count the amount of ['|"] in each array element...create a new array to hold it
For $i = 1 to UBound ( $atest ) - 1
 ReDim $asHolder[$i][2]
 $iCount = 0
 $iStart = StringRegExp (  $atest[$i], "['""]", 3 )
 If IsArray ( $iStart ) Then
  $iCount = UBound ( $iStart )
 EndIf
 ConsoleWrite ( $iCount & @CRLF )
 $asHolder[$i-1][0] = $atest[$i]
 $asHolder[$i-1][1] = $iCount
Next
_ArrayDisplay ($asHolder )
;;;would then add code to logically re-construct pieces of the above array
Dim $aFinal[1]
$iCurrent = 0
$iPrevQuoteCounts = 0
$iQuoteCount = 0
$iLast = 0
$sString = ""
$bSkipLogical = False
Dim $asComplete[1]
For $i = 0 To UBound ( $asHolder ) - 1
 ReDim $asComplete[$iCurrent+1]
 $iQuoteCount = $asHolder[$i][1]
 If StringLen ( $sString ) = 0 Then
  $sString = $asHolder[$i][0]
 Else
  $sString = $sString & $asHolder[$i][0]
 EndIf
 If IsInt (($iQuoteCount + $iPrevQuoteCounts)/2) Then
  $iPrevQuoteCounts = 0 ; Reset
  $asComplete[$iCurrent] = $sString
  $iCurrent = $iCurrent  + 1
  $sString = "" ; Reset
 Else
  $sString = $sString & " Or "
  $iPrevQuoteCounts = $iPrevQuoteCounts + $iQuoteCount
 EndIf
Next
_ArrayDisplay ( $asComplete )

Thanks, for any assistance.

Edited May 24, 2012 by jdelaney

[jchd]

Here's a sketch showing how you can do that. Sorry about the empty captures, I sincerely don't know how to avoid them!

#include <Array.au3>
; split by all or's
Local $atest = "@testand='and or string' and .='and or string2' or contains(.='test or quote') and @testand='or string2'"
Local $aCond = StringRegExp($atest, "(?ix) (?(DEFINE) (?<string> '[^']*') (?<var> . | @[a-z]w*) (?<equality> (?&var)=(?&string)) (?<expr> (?&equality) | contains((?&equality)))) ((?&expr)(?:s and s(?&expr))+) s or s ((?&expr)(?:s and s(?&expr))+)", 1)
For $i = 1 To 4
_ArrayDelete($aCond, 0)
Next
_ArrayDisplay($aCond)

Depending on the allowed input syntax, you may have to add whitespace metacharacters (like s* or [[:blank:]]* at some places to match all possible inputs.

[iamtheky]

This works for the above example with single quotes, might be another route...

#Include <Array.au3>
$string = "@testand='and or string' and .='and or string2' or contains(.='test or quote') and @testand='or string2'"
$Acapture = stringsplit($string , "' or " , 3)
for $i = 0 to ubound ($Acapture) - 2
$Acapture[$i] = $Acapture[$i] & "'"
next
_ArrayDisplay ($Acapture)

[jdelaney]

@jchd That works for the string provided, but not when I modify the string, such as changeing one of the and conditions to an or, it doesn't return the strings.

Thanks for the example though, I'm going to work on disecting it now.

Edited May 25, 2012 by jdelaney

[Robjong]

Hi,

What is the escape character? i.e. What character do you put in front of a quote if you want to put a single quote in a quoted sub string? ("@testand='input quote here: ' between quotes' or ...")

[jdelaney]

If you put two " in a surrounded quote, that is read as a single "...is that what you are asking?

"test and somestring""continued"

that is the same as a string, that is: test and somestring"continued

Edited May 25, 2012 by jdelaney

[GEOSoft]

The escape character is but look in the help file for x. in the case of the double quote (Chr(34)) you would use x22. In most cases where it could be either a double quote or a single quote you will be safe just using the class [:punct:] but in some situations that can produce unpredictable results.

There is a PCRE tool in my signature that will allow you to test against an actual file (Local file tab).

[jdelaney]

well, i made it ('it' being my code...much more simplified) much better, but still not fully getting how to do recursive regular expressions...so i did a loop instead

#include <Array.au3>
; split by all or's
Local $atest = "@testand='and or string' and .='and or string2' or contains(.='test or quote') and @testand='or string2' or @test='value' and test='orvalue or '"

$iCounter = 0
Dim $aHolder[1]
$aCond = StringRegExp($atest, "(?i)(?s)(.*'.*')sors", 3)
While IsArray ( $aCond )
ReDim $aHolder[$iCounter+1]
$temp = StringRegExp($atest, "(?i)(?s).*'.*'sors(.*)", 3)
If Not IsArray ( $temp ) Then
  $aHolder[$iCounter] = $aCond[0]
  ExitLoop
Else
  $aHolder[$iCounter] = $temp[0]
EndIf
$aCond = StringRegExp($atest, "(?i)(?s)(.*'.*')sors", 3)
$atest = $aCond[0]
$iCounter = $iCounter + 1
WEnd
_ArrayDisplay ( $aHolder)

Edited May 25, 2012 by jdelaney

[iamtheky]

i believe that, and my stringsplit example return the same array. Am i missing something as to the want for a regular expression?

Edited May 25, 2012 by boththose

[Robjong]

@jdelaney Yes, that's what I meant, thanks.

[jdelaney]

there may be a condition where there is a ') or ... scenario. Other then that, yours works... here is my update to consider that:

#include <Array.au3>
; split by all or's
Local $atest = "@testand='and or string' and .='and or string2' or contains(.='test or quote') or @testand='or string2' or @test='value' and test=' orvalue or '"

$iCounter = 0
Dim $aHolder[1]
$sBefore = "(?i)(?s)(.*'.*['[:punct:]])sors"
$sAfter = "(?i)(?s).*'.*['[:punct:]]sors(.*)"
$aCond = StringRegExp($atest, $sBefore, 3)
While IsArray ( $aCond )
 ReDim $aHolder[$iCounter+1]
 $temp = StringRegExp($atest, $sAfter, 3)
 If Not IsArray ( $temp ) Then
  $aHolder[$iCounter] = $aCond[0]
  ExitLoop
 Else
  $aHolder[$iCounter] = $temp[0]
 EndIf
 $aCond = StringRegExp($atest, $sBefore, 3)
 $atest = $aCond[0]
 $iCounter = $iCounter + 1
WEnd
_ArrayDisplay ( $aHolder)

[jdelaney]

Ok, this is the real deal...

#include <Array.au3>
; split by all or's
Local $atest = "@testand='and or string' and .='and or string2' or contains(.='test or quote') or @testand='or string2' or @test='value' and test=' orvalue '"
; split out ALL or conditions
#include <Array.au3>
$sRegExpOrSplit = "(?i)(?U)(.*'.*[')])(?:sors)|.{1,}?"; experiment combining
$aCond = StringRegExp($atest, $sRegExpOrSplit, 3)
_ArrayDisplay ( $aCond )

Edited May 25, 2012 by jdelaney

[jchd]

This modified version works as expected on your new example.

#include <Array.au3>
; split by all or's conditions
Local $atest = "@testand='and or string' and .='and or string2' or contains(.='test or quote') or @testand='or string2' or @test='value' and test=' orvalue '"
Local $aCond = StringRegExp($atest, "(?ix) (?(DEFINE) (?<string> '[^']*') (?<var> . | @?[a-z]w*) (?<equality> (?&var)=(?&string)) (?<expr> (?&equality) | contains((?&equality))) (?<andCond> (?&expr)(?: s and s (?&expr) )*)) ((?&andCond)) (?= s or s |$)", 3)
For $i = UBound($aCond) - 1 To 0 Step -1
If Mod($i + 1, 6) Then _ArrayDelete($aCond, $i)
Next
_ArrayDisplay($aCond)

[UEZ]

Is this working too when you change some values?

#include <Array.au3>
Local $atest = "@testand='and or string' and .='and or string2' or contains(.='test or quote') and @testand='or string2'"
Local $sBool = "or"
Local $aCond = StringRegExp($atest, "(?i)(@.*')s+" & $sBool & "s+([a-z]+.*)", 3)
_ArrayDisplay($aCond)

Br,

UEZ

Edited May 26, 2012 by UEZ

[jdelaney]

@ UEZ, there may be a '@' following the or, so that only split a few times. @ jchd, i need a regexp that can parse out dynamically. Thanks for all the help though, this works (though can be simplified further):

$sRegExpOrSplit = "(?i)(?U)(.*'.*[')])(?:sors)|.{1,}?"