lucene-solr-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Kissue Kissue <kissue...@gmail.com>
Subject Re: DIH primary key
Date Mon, 05 Sep 2011 06:16:29 GMT
Thanks for replying. Unfortuately the table i need to import from is a view
and there is no unique key in there i can use as a primary key. How does
this affect my using DIH? does it mea i cannot use DIH?

Thanks.



On Sun, Sep 4, 2011 at 8:44 PM, Shawn Heisey <solr@elyograg.org> wrote:

> On 9/4/2011 12:16 PM, Kissue Kissue wrote:
>
>> I was reading about DIH on the this Wiki link :
>> http://wiki.apache.org/solr/**DataImportHandler#A_shorter_**data-config<http://wiki.apache.org/solr/DataImportHandler#A_shorter_data-config>
>> The following was said about entity primary key: "is *optional* and only
>> needed when using delta-imports". Does this mean that the primary key is
>> mandatory for delta imports? I am asking because i am going to be
>> importing
>> from a view with no primary key.
>>
>
> I believe what it means is that you have to specify a field to be the
> primary key, and that it must exist in all three queries that you defined -
> query, deltaQuery and deltaImportQuery.  In my case, query and
> deltaImportQuery are identical, and deltaQuery is "SELECT 1 AS did".  The
> only thing this query does is tell the DIH that there is something to do for
> a delta-import, which it then uses deltaImportQuery to do.  I keep track of
> which documents are new outside of Solr and pass values for the query in via
> the dataimport URL.
>
> As you might surmise, did is the primary key in my dataimport config file.
>  I couldn't say what would happen if your query results have duplicate
> values in the primary key field.  In my case, did actually is is the primary
> key in the database, but I don't think that's required.  I use different
> fields for primary key and uniqueKey.  This allows us a little extra
> flexibility in the index.
>
> Hopefully you do still have a field that is unique (even if it's not a
> primary key) that you can use as the primary key in your config file.  It's
> a good idea to have such a thing available to serve as the uniqueKey in
> schema.xml, for automatic overwrites (delete and reinsert) of documents that
> change.
>
> Thanks,
> Shawn
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message