[jira] [Commented] (LUCENE-4503) MoreLikeThis supports multiple index readers.
[ https://issues.apache.org/jira/browse/LUCENE-4503?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13483322#comment-13483322 ] Ying Andrews commented on LUCENE-4503: -- * Added support for multiple index readers so More Like This can generate a similary query based on multiple indexes. * This extends the MoreLikeThis feature to work with lucene MultSsearcher. * * For example: * Due to large size we may want to divide all sales index into: sales_1, sale_2, sales_3, ..., sales_n. * In this case we would best use parallel multi-searcher to do the search. Old MoreLikeThis.java doesn't support * this scenario. If the current document of interest comes from index sales_1, then the query returned from * like(int) and like(Reader, String) will only be based on index sales_1, which apparently does not reflect the * entirety of the whole document population. * * Modified: * constructors - MoreLikeThis(IndexReader), * MoreLikeThis(IndexReader, Similarity) * private method - createQueue(MapString, Int) * * Added: * constructors - MoreLikeThis(IndexReader, IndexReader[]), * MoreLikeThis(IndexReader, IndexReader[], Similarity) * * Notes: * When invoking method like(int) of this class, you have to pass in the NORMALIZED document number. * You can use the same algorithm used in lucene MultiSearcher class, specifically seen in * subSearcher(int) and subDoc(int) methods. MoreLikeThis supports multiple index readers. - Key: LUCENE-4503 URL: https://issues.apache.org/jira/browse/LUCENE-4503 Project: Lucene - Core Issue Type: Improvement Reporter: Ying Andrews Priority: Minor Labels: patch Attachments: MoreLikeThis.java.patch Original Estimate: 72h Remaining Estimate: 72h -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-4503) MoreLikeThis supports multiple index readers.
[ https://issues.apache.org/jira/browse/LUCENE-4503?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13483340#comment-13483340 ] Robert Muir commented on LUCENE-4503: - Can't you just pass a MultiReader instead? MoreLikeThis supports multiple index readers. - Key: LUCENE-4503 URL: https://issues.apache.org/jira/browse/LUCENE-4503 Project: Lucene - Core Issue Type: Improvement Reporter: Ying Andrews Priority: Minor Labels: patch Attachments: MoreLikeThis.java.patch Original Estimate: 72h Remaining Estimate: 72h -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-4503) MoreLikeThis supports multiple index readers.
[ https://issues.apache.org/jira/browse/LUCENE-4503?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13483486#comment-13483486 ] Ying Andrews commented on LUCENE-4503: -- Thanks for pointing it out, Robert. In the application I worked on, we had to support a mix of local and remote searchers. Due to the large scale and heterogeneous nature of our systems we had to be able to search anything that implements Searchable. We also had to take advantage of ParallelMultiSearcher to boost the performance. In a special case we had a ParallelMultiSearcher consisted of a group of local file indexes, a group of remote searchers whose data may come further from other remote searchers (kind like a tree) and one searcher that gets data from a SolrServer over the network. Therefore we had to adopt MultiSearcher instead of MultiReader strategy. We recently added MoreLikeThis feature into our heterogenous system. As you can see MultiReader is not an option in our environment. The links below roughly explains my situation. Thank you. http://lucene.472066.n3.nabble.com/MultiSearcher-vs-MultiReader-td546968.html http://mail-archives.apache.org/mod_mbox/lucene-java-user/200712.mbox/%3cof924d8f48.261c9541-onc22573a5.0077e70d-c22573a5.007a9...@il.ibm.com%3E MoreLikeThis supports multiple index readers. - Key: LUCENE-4503 URL: https://issues.apache.org/jira/browse/LUCENE-4503 Project: Lucene - Core Issue Type: Improvement Reporter: Ying Andrews Priority: Minor Labels: patch Attachments: MoreLikeThis.java.patch Original Estimate: 72h Remaining Estimate: 72h -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
