Re: [bdbxml] About BDBXML performance

On Jul 25, 2005, at 11:32 AM, Kiran Kumar M R wrote:

> Hi,
>
> I want to use BDBXML to store 2-3 MB XML file.
> The file will have XML data of structure  like
> <Level 1>
>     <Level 2>
>         <Level 3>
>               <Level 4>
>                    <Value> 100000<value>
>               </Level4>
>         </Level3>
>     </Level2>
> </Level1>
>
>
> My search key will be /Level1/Level2/Level3/Level4/value .

If that is your key, what information do you intend to retrieve?

> What will be the search time taken by BDB XML  on normal computer and
> server grade computer.

Are you storing just one file, or a number of them with similar
structure?

You will have to benchmark your specific searches.  Much
depends on more detailed structure of your document.

BDB XML 2.1.8 uses indexes to restrict searches to a set of
documents, but does not (yet) use indexes to navigate directly
to nodes within a document.

There is, however, an ability to index at the node level, and
use the XmlContainer::lookupIndex() method, which may
work for your application.  If so, that would be quite fast (a simple
index lookup).

> In this scenario what should be the ideal configration used for
> creation on container and document nodes.

You would need to use a node storage container.  Also, depending
on the number of documents, you may want a reasonably large
Berkeley DB cache.  However, if you have a single, 3MB document,
cache size won't be an issue as long as it's something like 10MB.

If you are storing multiple documents, or using the lookupIndex
method, you will need to add an index
on the node "Value" that may be either:
     node-element-equality-string
     node-element-equality-decimal

Depending on whether you want to treat the value as a string
or number.

Regards,

George

>
> Please help me at the earliest.
>
> Thanks,
> Kiran



------------------------------------------
To remove yourself from this list, send an
email to [email protected]