Regexpsubstitute doesn't work the way I need it to.

[Kotofanchik]

I needed to find numbers in the string variable. The first, second, third number among other symbols. And I noticed that the mask I'm searching for gives me not the first, but the last match. Is this somehow regulated? Maybe Regexpsubstitute finds the first match like a web parser? And not the last?
For now, I got out of the situation by turning the string backwards, which is not very convenient.

[ikarus1969]

Please post a piece of text you are searching for and the regular expression you use so far. It's not easy (at least for me) to really understand where the problem is. With the text you are searching and the regular expression you are using i can follow better.

edit: one idea is to use the option (?U), which means "ungreedy". you have to place that at the beginning of your regular expression. Maybe that helps; but anyway: please post text and the regular expression you are using so far.

[sl23]

Out of curiosity, still trying to learn the basics of regex, can you use (0-100) if searching for 3 digit numbers?

[Kotofanchik]

here is the text :z:z:160:z:z:z:460:z:
I need to capture the first digit 160 with one capture and 460 with another
Here is the expression

regexpsubstitute=1
substitute=".*:(\d+):.*":"\1"
I expect it to return 160, but it returns 460
That is, the last match. The input string itself can be much longer and have more digits, but I am only interested in the first two numbers.

I tried substitute="(?U).*:(\d+):.*":"\1"
I got a completely incomprehensible 160460z:
I guess that I added it incorrectly and (?U) worked as a capture, but I don't understand how to do it correctly

[balala]

here is the text :z:z:160:z:z:z:460:z:
I need to capture the first digit 160 with one capture and 460 with another
Here is the expression

regexpsubstitute=1
substitute=".*:(\d+):.*":"\1"
I expect it to return 160, but it returns 460

Try this substitution: Substitute="(?U)^.*:(\d+):.*$":"\1"

[Kotofanchik]

Try this substitution: Substitute="(?U)^.*:(\d+):.*$":"\1"

Yes, it works as it should, thank you

[balala]

Yes, it works as it should, thank you

You're welcome.

[Yincognito]

Just so it's clear why that happened: since ".*:(\d+):.*" doesn't have any kind of "anchor" at the beginning or end to know from where and to where to do the substitute, it does it everywhere (i.e. on each match) in the string, that's why the former attempt didn't succeed as desired. Generally speaking, it's almost mandatory to use some kind of anchor to "lock" the pattern to a certain area of the string if you need to substitute only one occurrence and not all, that's why balala's suggestion worked (that and the ?U ungreedy flag to match as few chars as possible with the * quantifier, of course).

By the way, it's not often used, but a ? after the first * quantifier, to make the pattern "^.*?:(\d+):.*$" has the same effect as using the ?U flag here. What ? does after a quantifier is to locally "invert the greediness" (or lack of, if ?U is used) set by the presence or absence of ?U. The ? acts locally where used, while ?U acts globally on the entire string.

[Yincognito]

As for "using (0-100) if searching for 3 digit numbers", the typical way is to use either \d\d\d or \d{3} for that, but of course that would match any 3 digit number, i.e. from 000 to 999.

Since regex can't "iterate" or "count" in a pattern other than counting how many times a pattern occurs, if only 0 to 100 with no zero padding is desired, a regex like (\d|[1-9]\d|100) to capture it, or (?:\d|[1-9]\d|100) to not capture it is more suited for that, since it would match all 1 digit numbers (i.e. 0...9) OR all 2 digits numbers that don't start with 0 (i.e. 10..99) OR 100. The | means the logical OR, while 1-9 means characters (and not numbers in the mathematical sense, since regex only works with strings and has no "understanding" of numbers) from 1 to 9 in the ASCII / Unicode table.