RegEx: step by step

Thomas 2020

before:
a cd eg g h ik
it should look like this:
A Cd Eg G H Ik

I want to do it step by step (Replace)
1 A cd eg g h ik
2 A Cd eg g h ik
3 A Cd Eg g h ik
4 A Cd Eg G h ik
5 A Cd Eg G H ik
6 A Cd Eg G H Ik

step by step: it works
a cd eg g h ik
(?<=^)(.)
\u$0
A cd eg g h ik

step by step: does not work
a cd eg g h ik
(?<=\s)(.)
\u$0
a cd eg g h ik

Where is the mistake?

Terry R

@Pan-Jan said in RegEx: step by step:

I want to do it step by step (Replace)

Firstly, use of the look behind can be problematic as you have seen. @guy038 had a recent post explaining it in some detail. Have a look at
https://community.notepad-plus-plus.org/topic/19006/regex-positive-look-behind-with/6
and maybe you can see the reason why. Otherwise since I’ve referred to @guy038, he may well pop his head in and explain again. From my perspective the look-behind is constrained by a fixed width, thus where ever the cursor is your look-behind ONLY looks back 1 character. It’s also explained at:
https://www.rexegg.com/regex-disambiguation.html

I looked at a different idea and my Replace function regex is:
Find What:(^|\s).
Replace With:\U$0\E

I’m not entirely happy with the replacement field, for 2 reasons. Firstly it does attempt to convert the zero width “start of line” or the space to uppercase. Of course that doesn’t cause an issue but I feel it’s bad programming. Secondly it didn’t work with the \u modifier (which is only for a single character), thus I used the \U and \E as replacements. In saying that, it does work in the single step replacement which is your original issue.

Terry

Thomas 2020

Thanks, big U … gets the job done
uppercase
\u = [[:upper:]]
\U = [^[:upper:]]

Why is there a bug here?

Terry R

I don’t think that is a bug. I believe it’s the issue I mentioned. Your look behind is NOT a fixed width. The ^ is a zero width “character”, start of line. The “e” is 1 character wide. That probably makes the regex invalid.

My regex solution definitely isn’t great. It was made in a hurry to test my idea. There is bound to be a much better solution for your example.

Terry

Thomas 2020

I already know.
(?<=^A.+M)XYZ are forbidden

I can answer in 20 minutes.
It’s not very smart and editing before 3 minutes too

Alan Kilborn

@Pan-Jan said in RegEx: step by step:

I can answer in 20 minutes.

You have to get a reputation score of 2+ before the 20 minute thing is removed for your account.
With the quality of your postings, I wish you good luck with that.

Terry R

@Terry-R said in RegEx: step by step:

I’m not entirely happy with the replacement field, for 2 reasons.

Now that I’ve had a bit more time to look at the question my final regex is:
Find What:\w+
Replace With:\u$0

As the \u will ONLY convert the first character of anything following it to upper case we don’t need to worry that $0 actually contains the entire “word”, not JUST the first character.

Terry