Re: Different behavior for q=goo.com vs q=@goo.com in queries?
Provide us what Analyzers/Filters are you using for Analysis of email field. - Grijesh -- View this message in context: http://lucene.472066.n3.nabble.com/Different-behavior-for-q-goo-com-vs-q-goo-com-in-queries-tp2168935p2183256.html Sent from the Solr - User mailing list archive at Nabble.com.
Re: Different behavior for q=goo.com vs q=@goo.com in queries?
On Fri, Dec 31, 2010 at 2:40 AM, mrw mikerobertsw...@gmail.com wrote: Basically, just what you've suggested. I did the field/query analysis piece with verbose output. Not entirely sure how to interpret the results, of course. Currently reading anything I can find on that. [...] From the above, it is not quite clear whether you followed the import of Erick's reply. Essentially, at this point, we are guessing in the dark, but Erick is very likely correct that this is an (unexpected) consequence of your analysis chain at index/query time. The admin/analysis page ought to show you how goo.com is indexed/queried vs. how @goo.com is indexed/queried. Alternatively, please show us the index/query analysis chain is configured in your schema.xml. Regards, Gora
Re: Different behavior for q=goo.com vs q=@goo.com in queries?
also try debugQuery=true and see why each result matched On Thu, Dec 30, 2010 at 4:10 PM, mrw mikerobertsw...@gmail.com wrote: Basically, just what you've suggested. I did the field/query analysis piece with verbose output. Not entirely sure how to interpret the results, of course. Currently reading anything I can find on that. Thanks Erick Erickson wrote: What steps have you taken to figure out whether the contents of your index are what you think? I suspect that the fields you're indexing aren't being analyzed/tokenized quite the way you expect either at query time or index time (or maybe both!). Take a look at the admin/analysis page for the field you're indexing the data into. If that doesn't shed any light on the problem, please paste in the fieldType definition for the field in question, maybe another set of eyes can see the issue. Best Erick -- View this message in context: http://lucene.472066.n3.nabble.com/Different-behavior-for-q-goo-com-vs-q-goo-com-in-queries-tp2168935p2169478.html Sent from the Solr - User mailing list archive at Nabble.com.
Different behavior for q=goo.com vs q=@goo.com in queries?
Using Lucid's Solr 1.4 distribution, if I index my email inbox and then search it by passing in different email expressions, I notice that I get different results based on whether the '@' character is included, even though the character is present in every email address in the field I'm searching. For example, q=goo.com returns multiple items, as expected. However, q...@goo.com return no results. Since every address containing goo.com also contains @goo.com, I would expect the same number of results. I get this from both the Solr admin console and from my application, which URL-encodes the query. I Googled, searched the Wiki, and grepped the Pugh and Lucid books, but don't see anything about this. Ideas? Thanks! -- View this message in context: http://lucene.472066.n3.nabble.com/Different-behavior-for-q-goo-com-vs-q-goo-com-in-queries-tp2168935p2168935.html Sent from the Solr - User mailing list archive at Nabble.com.
Re: Different behavior for q=goo.com vs q=@goo.com in queries?
What steps have you taken to figure out whether the contents of your index are what you think? I suspect that the fields you're indexing aren't being analyzed/tokenized quite the way you expect either at query time or index time (or maybe both!). Take a look at the admin/analysis page for the field you're indexing the data into. If that doesn't shed any light on the problem, please paste in the fieldType definition for the field in question, maybe another set of eyes can see the issue. Best Erick On Thu, Dec 30, 2010 at 2:11 PM, mrw mikerobertsw...@gmail.com wrote: Using Lucid's Solr 1.4 distribution, if I index my email inbox and then search it by passing in different email expressions, I notice that I get different results based on whether the '@' character is included, even though the character is present in every email address in the field I'm searching. For example, q=goo.com returns multiple items, as expected. However, q...@goo.com return no results. Since every address containing goo.com also contains @goo.com, I would expect the same number of results. I get this from both the Solr admin console and from my application, which URL-encodes the query. I Googled, searched the Wiki, and grepped the Pugh and Lucid books, but don't see anything about this. Ideas? Thanks! -- View this message in context: http://lucene.472066.n3.nabble.com/Different-behavior-for-q-goo-com-vs-q-goo-com-in-queries-tp2168935p2168935.html Sent from the Solr - User mailing list archive at Nabble.com.
Re: Different behavior for q=goo.com vs q=@goo.com in queries?
Basically, just what you've suggested. I did the field/query analysis piece with verbose output. Not entirely sure how to interpret the results, of course. Currently reading anything I can find on that. Thanks Erick Erickson wrote: What steps have you taken to figure out whether the contents of your index are what you think? I suspect that the fields you're indexing aren't being analyzed/tokenized quite the way you expect either at query time or index time (or maybe both!). Take a look at the admin/analysis page for the field you're indexing the data into. If that doesn't shed any light on the problem, please paste in the fieldType definition for the field in question, maybe another set of eyes can see the issue. Best Erick -- View this message in context: http://lucene.472066.n3.nabble.com/Different-behavior-for-q-goo-com-vs-q-goo-com-in-queries-tp2168935p2169478.html Sent from the Solr - User mailing list archive at Nabble.com.
