Are Punctuations Indexed?

Developers Query Processing
1

From Hipchat

[3:46 PM] LY: @here, quick question: what happen if I search like meta_name=“abc, def-xyz” ? and there is really an indexed document with meta name = “abc, def-xyz”
[3:47 PM] LB: it needs to be turned into a query operator
[3:47 PM] LB: if you have done it
[3:47 PM] LB: change search.html to search.json
[3:47 PM] LB: and find the queryAsProcessed
[3:49 PM] LY: the queryAsProcessed is changed to abc def xyz
[3:49 PM] LB: a little surpsised it is not
[3:49 PM] LB:
“abc def xyz”
[3:50 PM] LY: it is, sorry missing the quotation ;p
[3:50 PM] LB: a phrase query def must be after abc, xyz must be after def
[3:50 PM] LB: ah so it is
[3:50 PM] LB: yeah the QP will strip the weird chars
[3:50 PM] LY: is there any way we can search the result of name = “abc, def-xyz”?
[3:51 PM] LY: we cannot escape the comma, right?
[3:52 PM] LY: now we have a ticket, client wants to search a document with title contains, - ?
[3:52 PM] LB: thats not going to happen
[3:52 PM] LY: ok, thanks @LB
[3:53 PM] PL: It shouldn’t matter in practice because Funnelback makes those optimisations to the query to speed things up and punctuation is stripped. Having the punctuation in there doesn’t add anything useful
[3:55 PM] LY: yup, agree :slight_smile:

2

There is an experimental indexer option that may achieve what you are after.

Be warned though use of this is experimental and it’s use could have weird side effects.

…

-extra_idx_chars= - String specifies additional ASCII punctuation characters to be treated
as indexable in the same way as letters or digits. Note that many such characters
are subject to special rules which are not deactivated by listing them here.
The use of ampersand has been tested. Other characters on a suck-it-and-see basis.
Maximum string length is 10.