Solr Injection

[Ian Holsman <li...@holsman.net>]

Hi.

I've been playing with Kettle (http://kettle.pentaho.org/ ) as a method 
to inject data into Solr (and other things at the same time), and it 
looks really promising.

I was wondering if anyone else had some experience using it with Solr 
and if they set it up to add a document at a time, or wrote a single XML 
'add' document
and then added all of them in one lot

Ideally I would like to have Solr accept a REST style URL without all 
the XML bs around and just pass the fields in as parameters (which is 
alluded to in http://issues.apache.org/jira/browse/SOLR-85 )
and just pound the Solr master with lots of little posts when I do 
incremental updates for < 1000 things and use the CSV uploader for 
larger things.

Thoughts?

Re: Solr Injection

[Henrik Martin <he...@netgate.net>]

Take a look at the ROME API. They have support for Atom 0.3, 1.0, and
just about every RSS format on the planet. It's a very solid API and
widely used in the open source community. There was something on their
mailing list about support for the publishing part of Atom not too
long ago. I think there is an implementation of that as well.
Here's a link to ROME's wiki:

http://wiki.java.net/bin/view/Javawsxml/Rome

/Henrik

On Tuesday 03 July 2007 14:58, Ryan McKinley wrote:

> What would an update request look like?  How close/far is this from the
> existing format/functions?  Is it just a syntax change from what we have
> or would it require something more?
>
> Is this something that can be handled with XSLT processing an atom file?
> http://issues.apache.org/jira/browse/SOLR-285
>
> Jeff Rodenburg wrote:
> > On 7/3/07, Walter Underwood <wu...@netflix.com> wrote:
> >> The Atom Publishing Protocol would be a good choice for a rest API to
> >> Solr.
> >> That comes with a spec, interop testing, and an active community.
> >
> > +1

Re: Solr Injection

[Ryan McKinley <ry...@gmail.com>]

What would an update request look like?  How close/far is this from the 
existing format/functions?  Is it just a syntax change from what we have 
or would it require something more?

Is this something that can be handled with XSLT processing an atom file?
http://issues.apache.org/jira/browse/SOLR-285


Jeff Rodenburg wrote:
> On 7/3/07, Walter Underwood <wu...@netflix.com> wrote:
>>
>> The Atom Publishing Protocol would be a good choice for a rest API to
>> Solr.
>> That comes with a spec, interop testing, and an active community.
>>
>>
> +1
> 

Re: Solr Injection

[Jeff Rodenburg <je...@gmail.com>]

On 7/3/07, Walter Underwood <wu...@netflix.com> wrote:
>
> The Atom Publishing Protocol would be a good choice for a rest API to
> Solr.
> That comes with a spec, interop testing, and an active community.
>
>
+1

Re: Solr Injection

[Walter Underwood <wu...@netflix.com>]

The Atom Publishing Protocol would be a good choice for a rest API to Solr.
That comes with a spec, interop testing, and an active community.

wunder

On 7/2/07 6:22 PM, "Ian Holsman" <li...@holsman.net> wrote:

> Hi.
> 
> I've been playing with Kettle (http://kettle.pentaho.org/ ) as a method
> to inject data into Solr (and other things at the same time), and it
> looks really promising.
> 
> I was wondering if anyone else had some experience using it with Solr
> and if they set it up to add a document at a time, or wrote a single XML
> 'add' document
> and then added all of them in one lot
> 
> Ideally I would like to have Solr accept a REST style URL without all
> the XML bs around and just pass the fields in as parameters (which is
> alluded to in http://issues.apache.org/jira/browse/SOLR-85 )
> and just pound the Solr master with lots of little posts when I do
> incremental updates for < 1000 things and use the CSV uploader for
> larger things.
> 
> Thoughts?