Multiple replacement with StringRegExpReplace

[Faalamva]

Hello,

Could you please tell me what's wrong in the following instruction :

$Output = StringRegExpReplace($Input, "[àâäçéèêëîïñôöùûüÀÂÄÇÉÈÊËÎÏÑÔÖÙÛÜ]", "[aaaceeeeiinoouuuAAACEEEEIINOOUUU]")

I thought it would replace each character one by one, like the following sed command :

sed "y/[àâäçéèêëîïñôöùûüÀÂÄÇÉÈÊËÎÏÑÔÖÙÛÜ]/[aaaceeeeiinoouuuAAACEEEEIINOOUUU]/" Input.txt > Output.txt

But apparently, the destination regexp is taken as a whole.

For example, I get :

Déjà -> D[aaaceeeeiinoouuuAAACEEEEIINOOUUU]j[aaaceeeeiinoouuuAAACEEEEIINOOUUU]

instead of :

Déjà -> Deja

Is there a simple command for multiple regexp vs. regexp replacement ?

If no, I'll do some kind of loop to replace each characters one by one.

Cheers !

Edited February 18, 2013 by Faalamva

[Melba23]

Faalamva,

I believe you will need to run an SRER for each replacement letter - you can obviously group the letters to be replaced. I would also suggest a 2D array to hold the pattern strings and replacement letter - like this:

Global $aList[2][2] = [["[àâä]", "a"], ["[éèêë]", "e"]] ; obviously you include all of the possibilities

$sText = "Déjà"

For $i = 0 To UBound($aList) - 1
    $sText = StringRegExpReplace($sText, $aList[$i][0], $aList[$i][1])
Next

MsgBox(0, "Result", $sText)

Note that you can use a simple StringReplace for a single letter (c and n) - but if you using a loop it might be just as quick to run an regex on everything.

M23

[PhoenixXL]

You can even check the

Local $String = 'Déjà'

Local $Replaced

While 1
$Replaced = RegExBackRef($String, '(.*?)([àâäéèêë])(.*)', '\2', "ReplaceFunc")
If $Replaced = -1 Or @error Then ExitLoop
$String = $Replaced
WEnd

MsgBox(64, 'Test', $String)


Func ReplaceFunc($sMatch)
Switch $sMatch
Case "é","è","ê","ë"
Return "e"
Case "à","â","ä"
Return "a"
EndSwitch
EndFunc   ;==>ReplaceFunc

Added this as the third example

Regards

[Faalamva]

Melba23, that's pretty much what I have done, in a more elegant way, thank you !

PhoenixXL, thank you also for your example, though it looks a bit less straightforward.

Edited February 18, 2013 by Faalamva

[AZJIO]

$sText = 'DDDDDDDDeeeeeeeejjjjjjjaaaaaa'
$sText = StringRegExpReplace($sText, '(.)\1+', '\1')
MsgBox(0, 'Yes?', $sText)

[Faalamva]

Azjio, it doesn't answer this topic's problem, but by a funny twist of fate, your code will help me for another problem that will soon emerge : reduce all multiple consecutive characters. So, thank you too !

[czardas]

This is something I considered myself recently. It might be of use.

When making multiple string replacements within a single string, longer strings should generally be replaced first to avoid erroneously replacing substrings within subsequent search patterns. For this we need to sort the search patterns by length (in descending order).