We're currently only offering a single search box at Stanford; while
that is likely to change in the future, here's something everyone can
standard number searches can be added to the "default" keyword search
without degradation of other search results. We currently allow ISBN,
ISSN (i have to fix index to omit hyphen and to look for it in the
query) and the unique id of our records (the "ckey" in sirsi/unicorn
parlance). I will be adding OCLC and LCCN in the near future.
Here's what we did:
1. Get clean standard numbers into the index. I have code to share
for ISBN, ISSN, OCLC, LCCN. (None is needed for unique ids). My code
takes care of prefixes, suffixes, and very basic validation (up to a
2. Add the fields to the "default" search in web/sys/SOLR.php method
buildQuery() ... the default in the switch statement.
I added these at the bottom with a boost factor of 0.1. It's
sufficient to find matches on the numbers. Given the very low
likelihood that someone would search for numbers with so many digits,
degradation to search results is non-existent. That may change
slightly when I get ISSN searches working: nnnn-nnnn could be a user-
entered date range.
3. For extra leverage, don't have the UI code parse out 020, 022, 035
fields on the fly: use the standard number fields directly from the
respective fields SOLR responses. Note: these fields can have
multiple values. We've changed web/xsl/solr-convert.xsl and the UI
We're happy to share our code. The UI changes may take a bit, b/c
we've done extensive mods to the UI code, but I'm sure we can tease
the pertinent code out.