Need help with a GREP

[Doc Maik]

Hello!

I use a lot of cross-refs which consist of a chapter number and some text, separated by a space and everything enclosed in quotation marks. Example: "1.1 This is an example"

I made a GREP to find them:  (")(\d+.+?) (.+?)(")

For some reason, using the placeholders for left and right double quotation marks doesn't work in my document, but in another, so I used the "any" placeholder. So far, this GREP finds all cross-refs.

However, it's supposed to stop at the end, at the right double quotation mark that encloses the cross-ref.

Some of the headlines the cross-refs are made from also contain their own quotation marks, so that some of the cross-refs turn out as: "1.1 This is "another" example"

What happens when using the GREP? It would stop at the first right quotation mark. I have tried a lot, but I cannot find the correct extension for my GREP, to sort of skip both or all of the inner quotation marks. 

Anyone has a solution, other than to replace the double quotation marks in the headlines by single ones? This would of course solve the problem. Thanks in advance.

[Doc Maik]

As a nasty addition, in a French document the quotation marks are followed/led by an unconditional space, put by the translator: 

I also couldn't find a wildcard or placeholder that would include that unconditional space.  (") (\d+.+?) (.+?) (") and (".)(\d+.+?) (.+?)(.") don't work.

[Peter Kahrel · Accepted solution]

Those 'unconditional' spaces are in fact variable-width non-breaking spaces. You use ~S for them in a grep expression.

[Doc Maik]

Thanks. As I'm currently digging deeper into GREP than ever before, I have a lot to learn.

My thinking now was to OR the search to either find the quotation mark or the mark follow by that special space.

So I made this: (["]|["~S])(\d+.+?) (.+?)([~S"]|["])

It works, but for the French document it now wouldn't include the outer quote anymore. 

I actually don't see why, because only ["~S] and [~S"] fulfill the pattern.

Update: this seems to do it: ("|"~S)(\d+.+?) (.+?)(~S"|")

[Peter Kahrel]

I don't understand your "|" at all. The expression works, but I've no idea why and how.

To match the number, instead of (\d+.+?) it feels safer to use [\d.]+ 

Also, you need parentheses for grouping only if you reference the groups. Leaving them out when they're not needed makes the expression more readable. Here's a simpler version of your expression:

"~S[\d.]+ .+?~S"

>  I have a lot to learn.

It's worth all your trouble. There are various web sites with training material. For a comprehensive guide, see https://creativepro.com/now-available-grep-in-indesign-3rd-edition/

[Doc Maik]

I don't understand your "|" at all. The expression works, but I've no idea why and how.

By @Peter Kahrel

As far as I learned, everything after the | is used, until delimited. So it's, in words, "quotation mark OR quotation mark followed by unconditional space". 

[Peter Kahrel]

Oops! Of course! But in your "|"~S it means 'quotation mark OR quotation mark, followed by ~S', which doesn't make much sense and explains why the closing quote isn't found. Your [~S"]|["] says '~S or quote followed by quote'. Remember that [~S"] and ~S|" are equivalent.

[Doc Maik]

Remember that [~S"] and ~S|" are equivalent.

Hmm... I knew that, but had in mind the expression in brackets is AND, not OR.

[Peter Kahrel]

The character class is tested against single characters, which can only be one OR the other.

[Doc Maik]

To match the number, instead of (\d+.+?) it feels safer to use [\d.]+ 

I think that wouldn't include two- or three-digit numbers. 

[Peter Kahrel]

It would. It's a character class, saying 'a string consisting of digits and/or dots'.

[Peter Kahrel]

There is a grep modifier to find matching characters (brackets, quotation marks) but they won't help you here I don't think.So you'd probably need a script for things like "1.1 This is "another" example"