content stream/MLT

[Mike Anderson <mi...@mit.edu>]

I'm trying to understand how content stream works with respect to MLT. I did
a regular MLT query using a document ID and specifying two fields to do MLT
on and got back a set of results. I then copied the xml for the document
with the aforementioned ID and pasted it to a text file. Then I made the
query with stream.file=mlt_doc.xml, but my result set was completely
different and didn't really make sense.

Am I not using content streams correctly here? Or does solr not use the
schema when accepting a content stream?

Thanks in advance,
Mike

Re: content stream/MLT

[Erik Hatcher <er...@gmail.com>]

The MoreLikeThis content stream support is implemented such that the  
content stream is simply text, analyzed as if it was in the mlt.fl.   
It doesn't handle Solr XML as you'd expect - simply treats it as a  
string and analyzes it to get the terms out.

	Erik

On Dec 9, 2009, at 10:21 PM, Mike Anderson wrote:

> I'm trying to understand how content stream works with respect to  
> MLT. I did
> a regular MLT query using a document ID and specifying two fields to  
> do MLT
> on and got back a set of results. I then copied the xml for the  
> document
> with the aforementioned ID and pasted it to a text file. Then I made  
> the
> query with stream.file=mlt_doc.xml, but my result set was completely
> different and didn't really make sense.
>
> Am I not using content streams correctly here? Or does solr not use  
> the
> schema when accepting a content stream?
>
> Thanks in advance,
> Mike