Strangley enough, you can nest positive and negative lookaheads!

In steps: find All Caps with bu{2,}

followed by “[Ss]creen” (uppercase or lowercase 's'), we get

bu{2,}(?=s[Ss]creen)

(I'm assuming you want a space in between) but then not followed by “button”, which is (?!sbutton):

bu{2,}(?=s[Ss]creen(?!sbutton))

Note how the negative lookahead for “button” is inside the positive lookahead for “screen”. (The trick to insert in this editor is to insert them twice: \.)

Now for the bad news…

I would also like the Change to: format if someone knows.

Often asked, but unfortunately InDesign's GREP cannot automatically change uppercase to lowercase, or the reverse.

I've never worked with the ME version, but I bet you can find something like “Page Binding Side” in Document Setup. The Scripting guide shows this value can be either “Bound on the left” or “Bound on the right”, which nicely corresponds to a book that is read front-to-back or … (oh wait) I mean right-to-left (what you are most likely used to) or left-to-right (for Hebrew, Arabic, and probably a couple of other scripts).

(Correction) Um, it depends on the wording :)

If you consider how the pages turn while reading, an English book is “right to left'. But on second thought, maybe this can be best described as in which direction you read: left to right for English, right to left for Farsi.

.. That “oh wait” is of course because books in Farsi are also read front-to-back. It's just that the front is on the other side, the back side (which is then, of course, the front side).

I guess we need a bit more to go on, your code runs fine as is. If you need other labels for other contents, just keep repeating the last two lines.

Your image was a bit too small to recognize the problem, but this remark speaks volumes;

I'll get by by NOT using the Opentype fraction option ..

It seems you are applying this option to your entire text. Don't; just apply it to the numbers that are supposed to be formatted as fractions.

That's really interesting. I have some text such as “A:1″, where the number is a custom variable, and I can find it using

:~u

To find just the colon, I expected this to work:

: (?=~u)

It doesn't, so let's try the inverse: find a colon where there is no variable following!

: (?!~u)

This also matches nothing … and I'm totally sure one of these two ought to have found something!

InDesign's GREP is great (in general), but one of the most annoying featurettes is that it does not tell you why a particular search failed. A “bad” GREP expression gives the same result as a no-match-found: “Cannot find match”. But if a plain search finds nothing and the inverse (finding it “not”) also finds nothing, then you can be pretty sure ID just does not understand the expression.

Since lookahead does work correctly with other 'tilde' codes such as ~., I'm inclined to call out Bug, and you may very well report it to Adobe with a Bug Report.

“Find one by one” is actually a good way — of course, it depends on how you do that! The findText command is very reliable for these kind of operations, as it returns a complete list of all occurrences of your character style. The only caveat emptor is that the returned list of occurrences is a static copy; that is, if you change something in the text of the first found item, the references to all next occurrences will be off. The common solution is to work backwards through your found list of items (and findText can return the found list in reverse, so you don't even have to run 'backwards').

This is only necessary when you are doing something with the actual text, such as inserting or deleting characters, or adding a hyperlink, XML, anchored object or index marker — these all insert invisible codes into text. It's not necessary when changing formatting, such as adding underline or changing the font.

Here is a sample script that adds underline to all occurrences of a character style (just as a visible example; this could be done better by changing the character style definition!).

If speed is of the essence, you could venture into the realm of Advanced Scripting. One of InDesign's most powerful scripting commands is “everyItem”. What does this do? It gathers a list of every item of the kind you give a reference to; and it does so with an astonishing speed as well. When the list is gathered, you can inspect its various properties at leisure and decide what to do with each of the items.

everyItem(), and its often-partner getElements(), is a tough beast to master, and so I would not recommend this for starters :) It took Marc Autret not one but two blog articles to investigate it in full: On 'everyItem()' Part 1 and (no surprise) On 'everyItem()' Part 2. However, just to tease you, here is a variant of the script above that gathers every single formatted text span and then inspects each separate one for your applied character style.

Please be more specific:

Tried to change dictionary and spelling to swedish but it won´t work out for me.

What didn't work? How does one “change the dictionary” to Swedish? (A very common error is to 'change' the dictionary in Preferences, which changes just the preferences for that language, but does nothing else.)

Are you aware of the following index settings, called “Sort Options” in the Index panel menu?