Add a new name before and after names that start with a capital letter

El kohen Amal

HI,

I used the code you sent me, and I got the result on
the following text:

Myname1 several Myname2 Myname1 games Myname2 Myname1 like Myname2 Myname1 Sonic Myname2, Myname1 Mario Myname2, Myname1 Spiderman Myname2, Myname1 etc Myname2…
Myname1 Games Myname2 Myname1 by Myname2 Myname1 excel Myname2!

I wish that the change will be made only on words that contains at the beginning a capital letter like :

several games like Myname1 Sonic Myname2, Myname1 Mario Myname2, Myname1 Spiderman Myname2, etc.
Myname1 Games Myname2 by excel!

Thanks again

PeterJones

@El-kohen-Amal said in Add a new name before and after names that start with a capital letter:

I wish that the change will be made only on words that contains at the beginning a capital letter

That’s what @Terry-R’s regular expression will do… but only if you have “Match Case” enabled, or if you use (?-i) at the beginning of the regular expression:

without match case = finds 10 matches
with match case = find 4 matches:
without match case, but with (?-i) prefixing finds 4 matches:

Thus, to ignore the status of the “match case” setting, I would recommend (?-i)(\u\w+) as the Find What search expression. (The replacement is the same as @Terry-R showed; I was just using the Find > Count to show number of matches quickly.)

El kohen Amal

Thank you Terry R and PeterJones for your support!
The codes are working well!

El kohen Amal

Thanks again

Terry R

@El-kohen-Amal said in Add a new name before and after names that start with a capital letter:

I wish that the change will be made only on words that contains at the beginning a capital letter like :

Apologies. I had been playing with both the “match case” button and (?-i) on and off as I was getting some conflicting results. When I finally presented my solution I must have had the match case button still ticked and didn’t realise.

Thanks to @PeterJones for coming to the rescue.

Terry

Aside to @PeterJones and @guy038 This seems counterproductive if a meta character such as \u which should mean upper case ALSO needs the match case or (?-i) to make it work correctly. Am I missing an important concept or is this a known issue with the regex engine?

PeterJones

@Terry-R said in Add a new name before and after names that start with a capital letter:

Am I missing an important concept or is this a known issue with the regex engine?

My guess is it’s just a quirk of the regex engine. If you wrote [A-Z], you wouldn’t be surprised that case-insensitive match makes it also match a-z… but when you obscure it with \u, it might seem somehow different to an end-user, but to the developer of the engine, “uppercase is uppercase, whether it’s explicit or part of an escape”. Put another way, in my interpretation, \u is just a shorthand notation for something that really is effectively equivalent to [A-Z] (or its much larger Unicode equivalent) – but I don’t know whether the implementation makes it exactly equivalent to the described character set or not.

guy038

Hello, @terry-r, @peterjones and all,

Surprising ! I’d never really noticed this particularity !

And, indeed, whatever the Boost valid syntaxes, below :

[[:upper:]], [[:u:]], \p{upper}, \p{u}, \pu, \u, [A-Z], [\x41-\x5A], [\x{41}-\x{5A}] or [\x{0041}-\x{005A}]

they all select one uppercase or lowercase letter, if you don’t tick the Match case option AND don’t have any (?-i) modifier in regex !

I also tried these forms [@-\\[] and [\x40-\x5B], so one char before A till one char after Z. But, oddly, in that case, it just matches the @ and the [ characters !

Now, If you think about a consecutive range of characters, by code-point, you could suppose, at first sight, that the regex [A-Z] would match only upper-case letters, which is a non-discontiguous set of upper-case characters.

But, what about the Unicode block Latin Extended-A, for instance. It contains all these characters:

ĀāĂăĄąĆćĈĉĊċČčĎďĐđĒēĔĕĖėĘęĚěĜĝĞğĠġĢģĤĥĦħĨĩĪīĬĭĮįİıĲĳĴĵĶķĸĹĺĻļĽľĿŀŁłŃńŅņŇňŉŊŋŌōŎŏŐőŒœŔŕŖŗŘřŚśŜŝŞşŠšŢţŤťŦŧŨũŪūŬŭŮůŰűŲųŴŵŶŷŸŹźŻżŽžſ

where, anyway, the upper-case and lower-case letters are mixed. So the search for a contiguous range of letters, with the same case, seems impossible, as our N++ Boost library is not built with full Unicode support !

In conclusion, I would say that, whenever we need to distinguish the case of letters in regexes, either :

Tick the Match case option, in dialogs
Insert a (?-i) modifier within the regex

For instance, the search of 1 upper-case letter, at a time, inside the Latin Extended-A Unicode block above, works, as expected, with the regex (?-i)\u

To my mind, the Peter’s interpretation seems the most plausible !

Best Regards,

guy038

PeterJones

@guy038 said,

regex (?-i)\u

ooh, as I looked at that juxtaposition this time, I realized: if we move the character escape sequence inside the option-parentheses, then @Terry-R can have the “must be this case, no matter what options are elsewhere” expressions: use (?-i:\u) for must-be-upper-no-matter-what and (?-i:\l) for must-be-lower-no-matter-what. It’s more to write, and we’ll forget that just as often as we forget to use (?-is) at the beginning of all our regex help, but…

guy038

Hi, @terry-r, @peterjones and all,

Nice find Peter ;-)) Indeed, if we consider the contents of a non-capturing group (?-i:.............) :

Any regex literal letter, inside this non-capturing group, matches the letter, itself, with its exact case
Any regex \u syntax, inside this non-capturing group, matches one upper-case letter
Any regex \l syntax, inside this non-capturing group, matches one lower-case letter
Any regex [A-Z] syntax, inside this non-capturing group, matches one letter, between A and Z
Any regex [a-z] syntax, inside this non-capturing group, matches one letter, between a and z

And this, whatever a possible i modifier, used, before, in the overall regex and/or whatever the Match case option is ticked or not !

BR

guy038

Alan Kilborn

@PeterJones said in Add a new name before and after names that start with a capital letter:

use (?-is) at the beginning

My rule of thumb is becoming to type that, first thing after calling up the Find window, when I know I’m doing regex.

I can always modify it later, but it is a good starting point.
The discussion in this thread about case bears that out.