The question mark is indeed the cause of the malfunctioning.

By default, GREP is Greedy: the operators ?, +, and * will match as much as possible (right until the rest ofthe expression puts a stop on it). If an expression ends with one of these, it will continue matching as much as possible until the repeated expression no longer matches. A question mark changes this: after a ?, +, or * will make it match as *little* as possible.

So what happens on a string of digits, such as “abc123efg”? Using “\d+”, it will match the entire set of digits “123” — as long as possible. With the added question mark “\d+?” it will match *as little as possible* (while still being a valid expression), and so just “1” is already enough.

In the OP’s case this happens because the Repeat modifier is at the very end of the GREP expression. If there is anything in the epression *after* this Repeat modifier, such as “\d+?” (digits but with a next lowercase letter), then GREP has no other option than to match “123e”, because the shortest *possible* match MUST include all digits.

You get your original spaces back because ‘$0’ inserts the exact entire found text again. You can use parentheses to indicate a part of the search string, and then use ‘$1’ to insert the first group in the Replace field, ‘$2’ for the second, etc. — up to 9 groups. But of course you can only ‘recall’ a numbered group in Replace if you defined it with parentheses in the Search field (and I bet that’s why your own attempts with ‘$2’ failed).

A second change is you do not have to list all of your search strings separately; after all there has to be only ONE character, out of a certain limited set, between your spaces.

All together, you can search for this

([+=~=×÷])

and replace with

~S$1~S

There was a single ‘?’ in your search query, and checking the rest I suppose it’s a copy-paste problem and it should have been the ‘multiply’ operator × instead. I would also rather use a *real* mathematical minus ‘-‘ (U+2212) rather than an en-dash, because it aligns with the ‘+’ and is a bit thicker (then again, not all fonts contain it).

Stuart, do you mean something like https://jongware.mit.edu/idcs6js/pc_Page.html#documentOffset — “The sequential number of the page within the document”?

This is the index of the nth PHYSICAL page in the entire document, disregarding any and all page numbering. Since it is a Javascript index, it starts at 0, NOT at 1, as a “page number” would.

David, unless I am *very* mistaken the pattern is: “add 1 tab at the start of the line”.

(Oh wait, I am) Sorry; plus space-dash-space becomes tab-dash-tab.

That string (and most of the others in a Font object) is just the font name as text. These texts are read out of the font *file*, but the filename can be just about *anything*. For example, I have a Stempel Garamond: fontFamily “Stempel Garamond LT”, fontStyleName “Roman”, fullName “Stempel Garamond LT Roman” — but the file itself is called “lt50277.pfb”.

What you are looking for is ‘location’: https://jongware.mit.edu/idcs6js/pc_Font.html#location (“The full path to the Font”).

Ouch :-) I would love to, but my boss would surely object! I’ve got a regular job at a design office.

For crash courses, you can check out IDSecrets’ GREP section: https://creativepro.com/resources/grep

.. and one of the most recommended guides on Using GREP with InDesign happens to be written by a Dutchman as well: Peter Kahrel’s “GREP in InDesign” (https://shop.oreilly.com/product/9780596156015.do)

Still no problem ;-)

\([^)]*\u[^)]*\)

This looks for an opening parenthesis, then zero-or-more Not-Close-Paren, followed by an uppercase character, followed by zero-or-more Not-Close-Paren and finally followed by a closing parenthesis.

The Not-Close-Paren sequences are necessary because essentially you want to skip a list of “Not-A-Capital” — which is possible, but that also includes the closing parenthesis. In your example it would start at “(this one..” and end after the “A)”, picking up everything in between as well.

Possibly you made a mistake with your parentheses: both OPEN and CLOSE need to be escaped. This works for me:

\(\u\)