What is the deal with "space" in the search for a name?

DEVONthink Feedback
1

I’m trying to find records whose name contains “BSR 2021-”. Which is a no-brainer with name:BSR 2021-, I get both existing records just fine.

Now, I’m a careful guy and know that naming can be sloppy, so there might be more than one space in the file name, like “BSR 2021-” (that’s two spaces, though you can’t see them). Consequently, I use name:BSR*2021-. Which leads to no record being found. Whereas the documentation says, “* matches none, one, or multiple characters”. In my mind, two spaces are “multiple characters”, so “*” should match them. It doesn’t.

OTOH, name:BSR* matches all records containing “BSR”, so it does match the initial space (and everything else after that).

Also, name:BSR[ ]2021-, where [ ] is a character class containing exactly one character, namely a space, doesn’t match any records. But syntactically, [ ] is very much the same as a single space character.

Apparently, I’m missing something here. How would I specify a robust search for a name containing “BSR” followed by an arbitrary number of arbitrary characters, followed by “2021-”?

BTW, the ? (matches exactly one character) does not match the space, either. And using quotes around the expression doesn’t help, either.

2

Extraneous spaces should be stripped IIRC.

Copy the name and paste it into a text document.
Then enable File > Show Invisible Characters.

3

Ah, now I see. You’re saying that I cannot even use two consecutive spaces in a filename because DT collapses them to one. Indeed, it does.

However, my irritation about * matching space sometimes (i.e. in “BSR*” matching everything after BSR, including the first space) and sometimes not (i.e. in the middle of the search string “BSR*2021-”) remains. As well as not understanding why [ ] doesn’t match a single space (which it should, according to documentation).

4

Criss would have to comment further on the specifics (and perhaps I need to adjust the documentation as well).

5

Redundant white spaces are indeed trimmed (also unnecessary ones at the beginning/end)

The search index doesn’t include spaces or separators, therefore the phrase “BSR 2021” matches all cases.