The short form is that you must provide and identify a unique key (analogous to a database PK). See http://wiki.apache.org/solr/UniqueKey
<http://wiki.apache.org/solr/UniqueKey>There's an example here: http://wiki.apache.org/solr/DataImportHandler <http://wiki.apache.org/solr/DataImportHandler>But one red flag in your mail is that you're thinking in terms of tables. Think about *denormalizing* your data. There's really no such thing as a join in SOLR/Lucene, and attempts to emulate a DB-style join should be resisted <G>... I know, I know. It really goes against the trained instincts of a DB person to *replicate* data all over the place. But search engines and RDBMSs are two very different beasts and trying to make one behave like the other is usually...er... unfortunate. Before you get too far into your migration, I'd *strongly* advise you to spend some time imagining what form a query would take with your proposed schema. Don't even bother with using SOLR query syntax at first, just construct your query with boolean logic. Something like id:234 AND title:wonderful AND title:life. No sub-selects, joins, etc. allowed. This will inform your schema no end.... HTH Erick On Mon, Dec 7, 2009 at 9:13 AM, Lee Smith <l...@weblee.co.uk> wrote: > Hello All > > Sorry newbie Q. > > Im looking at using the Data Import Handler to add my data to solr. > > But I am a little confused how I go about updating the index. I understand > there is no update index so just a delete replace but how will solr know > what to remove and add ? > > Also hope someone does not mind giving me advice on my scema I should use. > > I will be indexing multiple tables as each table means a different type of > search. Here is the tables and the rows im looking at adding to solr. > > Files: > - id > - display_name > - server_path > - file_type > - project_id > > Folders: > - id > - folder_name > - fullpath > - project_id > > Dailies: > - id > - scene > - take > - description > - filename (join) > - project_id > > Assets > - id > - title > - project_id > > Calendar: (Events) > - id > - title > - description > - project_id > > On top of this I will be looking at doing full indexing using solr cell of > the documents held in the file data table. > > Hope some can point me in the right direction and thank you in advance > > Regards > > Lee > > >
