RegEx Help with Backreference I think

Robert Or Janet Diebel

@guy038 said in RegEx Help with Backreference I think:
I am struggling with trying to accomplish what I think is required.
FIND BookName and isolate into a Group
Indicate the text to be acted on
Input required change, the goal
Put the change into the REPLACE field

I think that is it, but fitting it into a lookahead RegEx is still eluding me.

I will continue to work on it, but I have been for hours, and any guidance would be great.

guy038

Hi, @robert-or-janet-diebel,

I think that the most simple solution is to verify that, throughout your file, any “space” character, after the verse number, is ALWAYS a tabulation char !

So, given your ssimple INPUT text :

Kulava 1
Kulava 1:1 Halikumbi lyakavanga mukakweji wamuchivali, mumwaka wamuchivali kufuma haze valovokele mulifuchi lyaEjipitu, Yehova alwezele Mose mumakango aShinai muMbalaka yakuliwanyina ngwenyi,
Kulava 1:2 Lavenu lizavu lyavana vaIsalele lyosena mwaya jitanga javo muvisaka vyavakakuluka javo mukuvula chamajina avo, malunga vosena umwe naumwe,
Kulava 1:3 kufuma kuli ava vamyaka makumi avali (20) nakusambula, vosena vaze vakuhasa kulwa jita muli vaIsalele. Ove naAlone muvalave halizavu halizavu.
Kulava 1:4 Kaha kutanga hitanga kufume lunga apwenga nayenu, kaha mutu himutu apwenga mwata wakutanga yavakakuluka jenyi.
Kulava 1:5 Awa akiko majina amalunga navamikafwa — mutanga yaLuvene mufume Elizule mwanaSheteule.
Kulava 1:6 Mutanga yaShimeyone mufume Shelumiyele mwanaZulishatai.
Kulava 1:7 Mutanga yaYuta mufume Nashone mwanaAminatave.
Kulava 1:8 Mutanga yaIsakale mufume Netanele mwanaZuwale.
Kulava 1:9 Mutanga yaZevulune mufume Eliyave mwanaHelone.
Kulava 1:10 Vamuli vana vaYosefwe navapwa ava — mutanga yaEfwalime mufume Elishama mwanaAmihute, mutanga yaManase mufume Ngamalyele mwanaPetazule.

The following regex S/R :

FIND ^(.+ \d+:\d+)[\x20\t]+
REPLACE ${1}\t

would return this OUTPUT text :

Kulava 1
Kulava 1:1	Halikumbi lyakavanga mukakweji wamuchivali, mumwaka wamuchivali kufuma haze valovokele mulifuchi lyaEjipitu, Yehova alwezele Mose mumakango aShinai muMbalaka yakuliwanyina ngwenyi,
Kulava 1:2	Lavenu lizavu lyavana vaIsalele lyosena mwaya jitanga javo muvisaka vyavakakuluka javo mukuvula chamajina avo, malunga vosena umwe naumwe,
Kulava 1:3	kufuma kuli ava vamyaka makumi avali (20) nakusambula, vosena vaze vakuhasa kulwa jita muli vaIsalele. Ove naAlone muvalave halizavu halizavu.
Kulava 1:4	Kaha kutanga hitanga kufume lunga apwenga nayenu, kaha mutu himutu apwenga mwata wakutanga yavakakuluka jenyi.
Kulava 1:5	Awa akiko majina amalunga navamikafwa — mutanga yaLuvene mufume Elizule mwanaSheteule.
Kulava 1:6	Mutanga yaShimeyone mufume Shelumiyele mwanaZulishatai.
Kulava 1:7	Mutanga yaYuta mufume Nashone mwanaAminatave.
Kulava 1:8	Mutanga yaIsakale mufume Netanele mwanaZuwale.
Kulava 1:9	Mutanga yaZevulune mufume Eliyave mwanaHelone.
Kulava 1:10	Vamuli vana vaYosefwe navapwa ava — mutanga yaEfwalime mufume Elishama mwanaAmihute, mutanga yaManase mufume Ngamalyele mwanaPetazule.

You could suppose that using a look-behind, for the first part of the whole search regex, would work :

FIND (?<=^.+ \d+:\d+)[\x20\t]+
REPLACE \t

But, this construction is illegal as our Boost regex engine does not support look-behinds of VARIABLE length !

To get around this drawback, we may use the \K feature :

First, the search regex matches any leading part ..... \d+:\d+
As soon as it meets the \K feature, the regex engine :
- Cancels any previous search, so far
- Adjusts the engine regex position to the position of \K so, right after the verse number and before any mixed range of tabulations or space characters ( [\x20\t]+ )
- Re-start the search of the final part [\x20\t]+

And leads to the following regex S/R :

FIND ^.+ \d+:\d+\K[\x20\t]+
REPLACE \t

IMPORTANT : If you’re using this second regex S/R with the \K feature, the replacement MUST be a global one, using the Replace All button. You CANNOT use the step by step replacement with the Replace button !

Best regards,

guy038

Robert Or Janet Diebel

@guy038 Thanx for your kind dedication and patience in helping me. I can now do in a couple of minutes what was taking days.
Starting again armed with the new MACROs is indeed a new beginning.
I wish you and Terry every blessings

guy038

Hello, @robert-or-janet-diebel,

Regarding my last regex to normalize “space” characters to ONE tabulation char, only, I thought of a better regex S/R, which should speed up all the process as it skips all the correct lines containing just one tabulation character !

My very-Best-Novel 20:10  Mutanga yaYuta mufume Nashone mwanaAminatave.
My very-Best-Novel 20:10		   Mutanga yaYuta mufume Nashone mwanaAminatave.
My very-Best-Novel 20:10	Mutanga yaYuta mufume Nashone mwanaAminatave.
My very-Best-Novel 20:10		Mutanga yaYuta mufume Nashone mwanaAminatave.
My very-Best-Novel 20:10 Mutanga yaYuta mufume Nashone mwanaAminatave.

FIND ^.+ \d+:\d+\K(?:[\x20\t]{2,}|\x20)
REPLACE \t
Click on the Replace All button

The third line is skipped as it just contains 1 tab char and we get the OUTPUT text :

My very-Best-Novel 20:10	Mutanga yaYuta mufume Nashone mwanaAminatave.
My very-Best-Novel 20:10	Mutanga yaYuta mufume Nashone mwanaAminatave.
My very-Best-Novel 20:10	Mutanga yaYuta mufume Nashone mwanaAminatave.
My very-Best-Novel 20:10	Mutanga yaYuta mufume Nashone mwanaAminatave.
My very-Best-Novel 20:10	Mutanga yaYuta mufume Nashone mwanaAminatave.

Note the use of a non-capturing group (?:......) as we do not need the contents of this group, either, in the search and/or the replace regex !

Best Regards,

guy038

Robert Or Janet Diebel

@guy038 Thank You very much for the extra effort.