lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gerard Sychay" <>
Subject Re: Searching against Database
Date Mon, 19 Jul 2004 15:17:38 GMT
You might run into problems with having too many Fields by treating each
record as a Document and each column as a Field in that Document. An
alternative would be to index each cell of the table as a Document and
store and keep metadata (primary key, column name, table name, etc.) as
stored, untokenized, unindexed Fields.  This depends on of course how
many columns are in your tables, but Lucene can obviously deal with a
LOT of documents.

As far as updating, our requirements allowed for nightly updates, so I
can't help you there.

>>> "Peter M Cipollone" <> 07/15/04 08:05PM >>>

> Is it possible to search against the column in the table ? If so are
> there any limitations on the # of columns one should target to
> against?

What you can search against all depends on how you index your columns. 
believe you mentioned that you had data in multiple tables for each
(or Document in Lucene).  If you map your your columns to Lucene
Fields, and
make sure that the primary key for each record is stored in the same
Document object as the columns (Fields), then you should be golden.

Someone earlier pointed out that Oracle allows Java in its stored
procedures, so if you use a single stored procedure to insert a new
that same procedure can create a matching Lucene Document and add it to

For updates, you will need to delete the Lucene document and then add a
copy of the updated record.  If you have a primary key field that is
and stored in Lucene, you can use

to delete the old version.


> any other suggestions?
> Thanks.
> -H
> To unsubscribe, e-mail: 
> For additional commands, e-mail:


To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message