# causes problems in facets

[James Fournie]

EG 2.0.5
PG 8.4

# / pound sign / hash sign / number sign causes problems in a facet field. This can be observed by:

a) select * FROM facet.field_entry where value ~ '#'
b) do a search for one of those and attempt to use the facet with the #

[Michael Peters (mrpeters)]

Confirmed.

Can be tested in Evergreen Indiana (http://evergreen.lib.in.us) with the series facet " The Nightmare club ; #1".

[Michael Peters (mrpeters)]

For what it's worth, I think James may have meant "metabib.series_field_entry" (or one of the other *_field_entry) tables, instead of facet.field_entry which doesn't appear to exist.

[Jason Stephenson (jstephenson)]

Needs confirmation in 2.1+.

[George Duimovich (george-duimovich)]

Confirmed in 2.2.0 system running OpenSRF 2.1.0 & PG 9.14.

[Dan Scott (denials)]

Looks like the query is getting converted into:

"series|seriestitle[Foolish # 1]"

instead of:

"series|seriestitle:[Foolish # 1]" (note additional colon)

[Dan Scott (denials)]

Huh. http://evergreen-ils.org/dokuwiki/doku.php?id=documentation:technical:search_grammar claims that's the right syntax (no colon). And it does work for other facets without a hash.

Oh, ding ding ding: per those docs, the hash mark is the "modifier marker"; the docs also claim that can be adjusted (presumably directly in OpenILS/Application/Storage/QueryParser.pm.? Not clear where else that assumption is hard-coded though).

I guess when it was defined, the thought was that it wouldn't appear in series identifiers...

[Jason Stephenson (jstephenson)]

Is this a problem with facets in general?

It is a problem for tpac?

If it only affects JSPAC, I'm inclined to set this to "won't fix."

[Dan Scott (denials)]

Yes, this is a problem with facets in general; the search grammar is applied to any interface, and can be confirmed by following Michael's link to the Evergreen Indiana catalog (now running TPAC) and searching for series "The Nightmare Club", then trying to narrow down to one of the facets.

[Dan Scott (denials)]

Pushed a branch (user/dbs/pound_facets) to add some "#" facets to concerto.sql to make this easier to reproduce. Search for "Piano concerto in C major, op. 39" or "Beethoven: concertos and overtures" and you'll get facets with # signs.

[Jeff Godin (jgodin)]

The working branch user/tsbere/qp_fix referenced in bug 856811 contains a change which moves from using # to separate facets to using ][ to separate facets.

[Jennifer Pringle (jpringle-u)]

This bug is no longer present on 2.4 in the TPAC.

[Remington Steed (rjs7)]

I am not able to reproduce the problem in 2.5.