Macro: Find and Replace maintaining case of the word being replaced

[paul11491]

I have a macro that finds words in one array and replaces them with corresponding ones in another. Essentially it is:

BritWords() = Array("fibre", "foetal")
	USWords() = 	Array("fiber", "fetal")
	Doc = ThisComponent
	Replace = Doc.createReplaceDescriptor
	 
	For I = LBound(BritWords()) To UBound(BritWords())
	  Replace.SearchString = BritWords(I)
	  Replace.ReplaceString = USWords(I)
	  Doc.replaceAll(Replace)
	Next I

It works great, but with the problem that the case of replacement word is imposed on the original word. So in the above case, if “Fibre” is at the beginning of a sentence, and thus is capitalized, it will be replaced with “fiber”, which is a usage error.

So I’m looking for a way to respect the case of the word being replaced. Would it be to capture the capitalization state and then impose it immediately after the replacement? Or is there a parameter that could be set for .doc.replaceAll(Replace) (perhaps involving Regex if no other way?) to carry the case through? Thanks.

[mikekaganski]

Unless you would use some smart library, you will have to do it manually:

1. Find a case-insensitive match.
2. Build a vector of “upper/lower case” flags for each character (or, if only the first character matters, then it is a sole “start with upper/lower case” flag).
3. Apply it to the replacement (for the simple “first character” case, it’s trivial; for the “each character” case, you would have to decide what to do with character count mismatch - or even with a case of different count of words between the match and replacement).
4. Do the replacement with the prepared result from #3.

[paul11491]

Thanks. I think you’re saying what I was trying to get at. I understand the concept, actually coding it is another matter.

[mikekaganski]

By the way, even the “trivial” cases that I mention above, aren’t completely trivial: you also need to consider cases where the found matches start in the middle of a word; what to do with the replacements with mixed or upper case, etc., etc.

[paul11491]

Yes, true. I did notice that the doc.replaceAll(Replace) catches strings, not just whole words. But even if it’s not perfect, doing it this way clears up a lot of problems.