Closed GoogleCodeExporter closed 9 years ago
Issue 82 has been merged into this issue.
Original comment by mchol...@gmail.com
on 29 Oct 2012 at 3:54
Your search will effectively run without the term +uri:"/" because the slash
"/" is not an indexed character. When using the default ELSA configuration,
Sphinx will treat non-alphanumeric characters as word breaks, so a search for
slash will not get any hits, because it will not be a keyword. So, the search
in your example just looked for tomcat logs with no referrer and a given srcip
from a given host and disregarded the URI of slash.
I will add code to at least warn if a term ends up being completely removed.
If you want to be able to search for a URI with just slash, you can go into
archive mode by switching the "Index" button to "Archive." The search will be
quite slow, however.
Another alternative is to edit the sphinx.conf file on each log node and add in
a blended_chars config statement to the "temporary" and "permanent" index
templates. Docs on the blended_chars config is available here:
http://sphinxsearch.com/docs/current.html#conf-blend-chars . However, a small
ELSA change will also be needed to enable that, so let me know if you want to
go that route as I'll have to make a small code update so you can configure
ELSA to parse the queries with those blended chars correctly.
Original comment by mchol...@gmail.com
on 29 Oct 2012 at 4:55
Warning added in rev 492.
Original comment by mchol...@gmail.com
on 29 Oct 2012 at 5:29
Original issue reported on code.google.com by
Laurel.W...@gmail.com
on 29 Oct 2012 at 3:43Attachments: