This is a slow way to do this; databases are capable of doing this join and feeding the results very efficiently.
The 'skipDoc' feature allows you to break out of the processing chain after the first query. It is used in the wikipedia example. http://wiki.apache.org/solr/DataImportHandler On Sat, Jun 12, 2010 at 6:37 PM, Holmes, Charles V. <chol...@mitre.org> wrote: > I'm putting together an entity. A simplified version of the database schema > is below. There is a 1-[0,1] relationship between Person and Address with > address_id being the nullable foreign key. If it makes any difference, I'm > using SQL Server 2005 on the backend. > > Person [id (pk), name, address_id (fk)] > Address [id (pk), zipcode] > > My data config looks like the one below. This naturally fails when the > address_id is null since the query ends up being "select * from user.address > where id = ". > > <entity name="person" > Query="select * from user.person"> > <entity name="address" > Query="select * from user.address where id = ${person.address_id}" > </entity> > </entity> > > I've worked around it by using a config like this one. However, this makes > the queries quite complex for some of my larger joins. > > <entity name="person" > Query="select * from user.person"> > <entity name="address" > Query="select * from user.address where id = (select address_id from > user.person where id = ${person.id})"> > </entity> > </entity> > > Is there a cleaner / better way of handling these type of relationships? > I've also tried to specify a default in the Solr schema, but that seems to > only work after all the data is indexed which makes sense but surprised me > initially. BTW, thanks for the great DIH tutorial on the wiki! > > Thanks! > Charles > -- Lance Norskog goks...@gmail.com
