tag:blogger.com,1999:blog-13606853.post114820356651898222..comments2023-12-28T16:01:32.344+05:30Comments on Whatever....: mysql Fulltext search versus lucenegamegeekhttp://www.blogger.com/profile/00557771372755358783noreply@blogger.comBlogger19125tag:blogger.com,1999:blog-13606853.post-63174743801681785812010-04-09T12:13:22.520+05:302010-04-09T12:13:22.520+05:30To connect lucene to mysql, you will have to write...To connect lucene to mysql, you will have to write your own jdbc connector - connect from java to mysql and pull data from mysql to lucene.gamegeekhttps://www.blogger.com/profile/00557771372755358783noreply@blogger.comtag:blogger.com,1999:blog-13606853.post-36341502663520393162010-04-02T01:12:20.852+05:302010-04-02T01:12:20.852+05:30It's what I really matter, do you have some ex...It's what I really matter, do you have some example code? can you explain how to connect lucene and Mysql? I've just begun to read about lucene and I think it's perferct for my paper. I cannot imagine how lucene treats mysql tables. Thanks in advanceAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-13606853.post-18026798365285862282010-03-08T17:17:32.183+05:302010-03-08T17:17:32.183+05:30This comment has been removed by a blog administrator.Chetanhttps://www.blogger.com/profile/04085930731302404957noreply@blogger.comtag:blogger.com,1999:blog-13606853.post-33689861590337619182009-10-17T16:08:11.317+05:302009-10-17T16:08:11.317+05:30Thank you for the hint with Sphinx - I will have a...Thank you for the hint with Sphinx - I will have a look at this also.<br /><br />If performance will become an issue - I don't know. In general I prefer to do things well right from the beginning. Doing major changes later sucks.Martin Wildamhttps://www.blogger.com/profile/10078822365635360301noreply@blogger.comtag:blogger.com,1999:blog-13606853.post-30461193487819254832009-10-17T01:31:02.466+05:302009-10-17T01:31:02.466+05:30@Martin: If you want to combine a fulltext search ...@Martin: If you want to combine a fulltext search into a single query with relational tables, then mysql fulltext is the best option. But there is sphinx engine - also available, which could be plugged into mysql. It is much faster than default mysql-fulltext engine. <br /><br />Again, additional effort is only required when performance becomes a concern.gamegeekhttps://www.blogger.com/profile/00557771372755358783noreply@blogger.comtag:blogger.com,1999:blog-13606853.post-7676587550200275002009-10-16T16:47:11.530+05:302009-10-16T16:47:11.530+05:30I would really prefer to just use MySQL because I ...I would really prefer to just use MySQL because I am planning a project that should have a lot of relational information that should also be searchable accordingly. But I have large chunks of text (document content) that also needs to be indexed.<br /><br />So if I need to use Lucene I would miss the option to combine relational with fulltext search getting a single result.<br /><br />And Martin Wildamhttps://www.blogger.com/profile/10078822365635360301noreply@blogger.comtag:blogger.com,1999:blog-13606853.post-6467199534903156782008-08-01T14:39:00.000+05:302008-08-01T14:39:00.000+05:30Speed of sorting in Mysql depends on indexing! If ...Speed of sorting in Mysql depends on indexing! If you sort according to a column with a B-tree index, sorting is lightning fast, since mysql won't need to copy the data, only the index.Harald Grovenhttps://www.blogger.com/profile/12261585095514038197noreply@blogger.comtag:blogger.com,1999:blog-13606853.post-29999683261109448452007-10-19T07:33:00.000+05:302007-10-19T07:33:00.000+05:30This is a very helpful and informative comparison....This is a very helpful and informative comparison.<BR/>But specifically, what's the fundamental reason that makes Lucene is faster than database index, since database index may also be Hash table/map? Is it because of the segment strategy Lucene uses?Zixinhttps://www.blogger.com/profile/17793211804449557798noreply@blogger.comtag:blogger.com,1999:blog-13606853.post-42159041575737953032007-06-29T10:01:00.000+05:302007-06-29T10:01:00.000+05:30That was a very good comparison and was very usefu...That was a very good comparison and was very usefull<BR/><BR/>-Srinivas VarmaUnknownhttps://www.blogger.com/profile/17619882734576470967noreply@blogger.comtag:blogger.com,1999:blog-13606853.post-29683993880654384742007-06-20T08:29:00.000+05:302007-06-20T08:29:00.000+05:30Yup razvan, you are right. Mysql is meant for data...Yup razvan, you are right. Mysql is meant for data storage and retrieval but it also has fulltext indexing capabilities. What i have done is a simple comparison of mysql fulltext capabilities with the fulltext indexing capabilities of lucene.<BR/><BR/>But mysql was not meant for fulltext searches - as data storage and retrieval is its prime objective. Whereas in case of lucene, fulltext search isgamegeekhttps://www.blogger.com/profile/00557771372755358783noreply@blogger.comtag:blogger.com,1999:blog-13606853.post-79940479226098590702007-06-19T23:19:00.000+05:302007-06-19T23:19:00.000+05:30But why you want to compare a information retrieva...But why you want to compare a information retrieval library (lucene) with a relational database management system (MySQL) in the first place, when each of them is made with a different scope?<BR/>One is a specialised information retrieval tool that allows you to optimize full-text searching to your custom demands (and thus fast) while the other is a relational database (full-text searching is notAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-13606853.post-1149220716478084552006-06-02T09:28:00.000+05:302006-06-02T09:28:00.000+05:30lucene can index any type of textual data whether ...lucene can index any type of textual data whether it resides in mysql or in files or in any other database. <BR/><BR/>You will have to write down programs to fetch data from any data source be it mysql or db2 or text files and feed it into lucene to make the index.gamegeekhttps://www.blogger.com/profile/00557771372755358783noreply@blogger.comtag:blogger.com,1999:blog-13606853.post-1149204563544808582006-06-02T04:59:00.000+05:302006-06-02T04:59:00.000+05:30Can Lucene index the text content which resides in...Can Lucene index the text content which resides inside Mysql MyIsam tables, or "just" files?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-13606853.post-1148454679528850842006-05-24T12:41:00.000+05:302006-05-24T12:41:00.000+05:30Good overall comparison of Lucene vs Mysql fulltex...Good overall comparison of Lucene vs Mysql fulltext search!<BR/><BR/>CLucene is not in C. It's an implementation of Lucene in C++. Apache foundation started an index-level compatible project in C for Lucene, codename lucene4c, but it hasn't got very far as CLucene.<BR/><BR/>Another great indexing toolkit which is very much in the same ballpark as Lucene but less fanfare exposure is Xapian http://Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-13606853.post-1148318948190322752006-05-22T22:59:00.000+05:302006-05-22T22:59:00.000+05:30As far as i know CLucene is just a replication of ...As far as i know CLucene is just a replication of lucene-java version in c. The way indexing and search works for CLucene and for java-lucene is the same. <BR/><BR/>However what i have read is that CLucene <BR/>is 3 times faster than java-lucene due to the fact that the C code is closer to machine language without any intermediate virtual machine as in java.gamegeekhttps://www.blogger.com/profile/00557771372755358783noreply@blogger.comtag:blogger.com,1999:blog-13606853.post-1148309569499281062006-05-22T20:22:00.000+05:302006-05-22T20:22:00.000+05:30Has anyone considered placing CLucene (http://sour...Has anyone considered placing CLucene (http://sourceforge.net/projects/clucene/) in an UDF, it's index in a blob field and using it to perform full text searches in MySQL? Or am I going the wrong path?Anonymoushttps://www.blogger.com/profile/13441809988487585009noreply@blogger.comtag:blogger.com,1999:blog-13606853.post-1148270792892917492006-05-22T09:36:00.000+05:302006-05-22T09:36:00.000+05:30I will be putting down some figures to the perform...I will be putting down some figures to the performance benchmarks soon...gamegeekhttps://www.blogger.com/profile/00557771372755358783noreply@blogger.comtag:blogger.com,1999:blog-13606853.post-1148260956768340022006-05-22T06:52:00.000+05:302006-05-22T06:52:00.000+05:30Thanks Jayant. This will make a great explaination...Thanks Jayant. This will make a great explaination for that guy who asked on freenode#apache today about the differences. Let's hope he comes back again!Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-13606853.post-1148239931974075072006-05-22T01:02:00.000+05:302006-05-22T01:02:00.000+05:30Thanks for this very insightful comparison!It was ...Thanks for this very insightful comparison!<BR/><BR/>It was a pleasant read, and I think it gave me a graps of the differences between these two methods of fulltext indexing.<BR/><BR/>I've got a queston though, in a couple of places, you're telling something about performance issues: inserting new documents, the actual search and the sorting. Can you back this by a few figures? I'd be very Anonymoushttps://www.blogger.com/profile/13441809988487585009noreply@blogger.com