Jeff,

This makes sense. I can see how it would be difficult and messy to map
keys to locations within a file, especially since the locations can
change each time the document is modified.

I'll have to restructure things.

Thanks for your help.

Regards,
Dominic

On Sun, 2002-11-17 at 04:31, Jeff Greif wrote:
> According to my reading of the source code (in xindice 1.0), this should
> *not* work.  The indexes are over the collection, mapping the values of the
> indexed elements or attributes to document keys.  The index is used to find
> the document satisfying the constraints.  Then xpath is used to find the
> part of the document that is relevant.  In a collection containing just one
> document, the index is of little or no value.  That's why xindice is
> supposed to be better for many small documents than a few large ones.
> 
> Jeff
> 
> ----- Original Message -----
> From: "Dominic Gamble" <[EMAIL PROTECTED]>
> To: <[email protected]>
> Sent: Saturday, November 16, 2002 6:58 AM
> Subject: indexing within a single document?
> 
> 
> | Hi there.
> |
> | I'm trying to get some performance improvements by adding an index to a
> | collection which contains 1 large document.
> |
> | I've set it up in this order
> |
> | 1. Create the collection
> | 2. Add the index
> | 3. Add the document
> |
> | I'm sure i done it correctly, but I don't get any performance
> | improvement after indexing. I've done this with a large document (5mb)
> | so I expect the speed increase to be significant.
> |
> | Does indexing only help if the data is spread accross many many
> | documents?
> |
> | Is it possible to index within a single document, so that xpath queries
> | on that document using the nodes that are indexed, will be faster?
> |
> | Regards,
> | Dominic Gamble
> |
> |
> |
> 


Reply via email to