[jira] [Commented] (LUCENE-6350) switch TermsQuery to prefixcodedterms
[ https://issues.apache.org/jira/browse/LUCENE-6350?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14532464#comment-14532464 ] ASF subversion and git services commented on LUCENE-6350: - Commit 1678167 from [~jpountz] in branch 'dev/branches/branch_5x' [ https://svn.apache.org/r1678167 ] LUCENE-6350: TermsQuery is now compressed with PrefixCodedTerms. switch TermsQuery to prefixcodedterms - Key: LUCENE-6350 URL: https://issues.apache.org/jira/browse/LUCENE-6350 Project: Lucene - Core Issue Type: Task Reporter: Robert Muir Attachments: LUCENE-6350.patch, LUCENE-6350.patch, LUCENE-6350.patch This will save ram and cleanup a lot of the code. Unfortunately the code is still a mess, it has a custom iterator api, and prefixcodedterms has yet another custom iterator api (seriously, maybe the worst one ever). -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-6350) switch TermsQuery to prefixcodedterms
[ https://issues.apache.org/jira/browse/LUCENE-6350?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14532445#comment-14532445 ] ASF subversion and git services commented on LUCENE-6350: - Commit 1678164 from [~jpountz] in branch 'dev/trunk' [ https://svn.apache.org/r1678164 ] LUCENE-6350: TermsQuery is now compressed with PrefixCodedTerms. switch TermsQuery to prefixcodedterms - Key: LUCENE-6350 URL: https://issues.apache.org/jira/browse/LUCENE-6350 Project: Lucene - Core Issue Type: Task Reporter: Robert Muir Attachments: LUCENE-6350.patch, LUCENE-6350.patch, LUCENE-6350.patch This will save ram and cleanup a lot of the code. Unfortunately the code is still a mess, it has a custom iterator api, and prefixcodedterms has yet another custom iterator api (seriously, maybe the worst one ever). -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-6350) switch TermsQuery to prefixcodedterms
[ https://issues.apache.org/jira/browse/LUCENE-6350?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14532579#comment-14532579 ] David Smiley commented on LUCENE-6350: -- This PrefixCodedTerms utility is new to me -- pretty cool. It seems similar to an FST configured to be an FSA. Have they been benchmarked between each other before? I would expect the FST/FSA to compress better. switch TermsQuery to prefixcodedterms - Key: LUCENE-6350 URL: https://issues.apache.org/jira/browse/LUCENE-6350 Project: Lucene - Core Issue Type: Task Reporter: Robert Muir Fix For: Trunk, 5.2 Attachments: LUCENE-6350.patch, LUCENE-6350.patch, LUCENE-6350.patch This will save ram and cleanup a lot of the code. Unfortunately the code is still a mess, it has a custom iterator api, and prefixcodedterms has yet another custom iterator api (seriously, maybe the worst one ever). -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-6350) switch TermsQuery to prefixcodedterms
[ https://issues.apache.org/jira/browse/LUCENE-6350?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14531488#comment-14531488 ] Michael McCandless commented on LUCENE-6350: +1, thanks [~jpountz]! switch TermsQuery to prefixcodedterms - Key: LUCENE-6350 URL: https://issues.apache.org/jira/browse/LUCENE-6350 Project: Lucene - Core Issue Type: Task Reporter: Robert Muir Attachments: LUCENE-6350.patch, LUCENE-6350.patch, LUCENE-6350.patch This will save ram and cleanup a lot of the code. Unfortunately the code is still a mess, it has a custom iterator api, and prefixcodedterms has yet another custom iterator api (seriously, maybe the worst one ever). -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-6350) switch TermsQuery to prefixcodedterms
[ https://issues.apache.org/jira/browse/LUCENE-6350?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14372617#comment-14372617 ] Michael McCandless commented on LUCENE-6350: I think we should commit LUCENE-6315 (simplifies this package-private iterator) before making this iterator public? Also, can you mark PrefixCodedTerms @internal (NOT @experimental) before making it public? This is not intended to be a publicly consumable API. You don't need to do the field.equals(lastField) on each term? The iterator tells you when it switches to a new field. You added an if (disi == null) check, which is great; do we have a test that tickles that? Is RAMFile.equals/hashCode or PrefixCodedTerms.equals/hashCode ever used anywhere else in the code base (because now they are quite costly)? I assume not... switch TermsQuery to prefixcodedterms - Key: LUCENE-6350 URL: https://issues.apache.org/jira/browse/LUCENE-6350 Project: Lucene - Core Issue Type: Task Reporter: Robert Muir Attachments: LUCENE-6350.patch, LUCENE-6350.patch This will save ram and cleanup a lot of the code. Unfortunately the code is still a mess, it has a custom iterator api, and prefixcodedterms has yet another custom iterator api (seriously, maybe the worst one ever). -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-6350) switch TermsQuery to prefixcodedterms
[ https://issues.apache.org/jira/browse/LUCENE-6350?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14355826#comment-14355826 ] Adrien Grand commented on LUCENE-6350: -- +1 switch TermsQuery to prefixcodedterms - Key: LUCENE-6350 URL: https://issues.apache.org/jira/browse/LUCENE-6350 Project: Lucene - Core Issue Type: Task Reporter: Robert Muir Attachments: LUCENE-6350.patch, LUCENE-6350.patch This will save ram and cleanup a lot of the code. Unfortunately the code is still a mess, it has a custom iterator api, and prefixcodedterms has yet another custom iterator api (seriously, maybe the worst one ever). -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
