xindice article on IBM's developerWorks

[David Viner <dv...@yahoo-inc.com>]

http://www-106.ibm.com/developerworks/library/wa-xindice.html

nothing major mentioned in the article, but it's cool to see Xindice being
written about by IBM.


dave

Re: Accessing documents in the database

[Terry Rosenbaum <Te...@amicas.com>]

Unfortunately, the current implementation of the Xindice
query result set is not incremental. The results
are collected into one large String in the server
before return to the requester. 138*292KB = about 40 MBytes.
That may be blowing memory available to the JVM.

Try using jvm commandline options to increase
available vm for your client and server (Xindice 1.0)
or for your single process if using embedded xindice
from CVS.

This issue is one that definitely requires attention in
future versions of xindice.

There is no way to query for the auto-assigned
document keys. A workaround might be to include your
own key element as a child of your ROWLIST element
e.g.

<ROWLIST>
    <K>your key here</K>
    <ROW></ROW>
    ...
</ROWLIST>

That way, you could query for "/ROWLIST/K"
and retrieve a much smaller result containing
the necessary keys. You could then query
for the documents in small batches (as small
as 1 document per query) using your own
keys to process the documents
one-by-one. (e.g.
"/ROWLIST[K='key1' or K='key2' or K='k3' or K='k4']"
to process 4 documents per pass).

-Terry

Satishkumar Ganesan wrote:

>Hi
>
>I have 138 documents each of 292 KB.
><ROWLIST> <ROW></ROW> </ROWLIST>
>I have 1000 nodes <ROW> </ROW> in each document.
>Now i need to do some processing in all the rows.
>The xpath query "/ROWLIST" doesn't return anything, this could be because of
>the numbers involved 138*100;
>Is there any method that would return all the document keys in the database,
>so that we can access one document at a time?
>
>
>Thanks
>Satish
>
>
>  
>

Accessing documents in the database

[Satishkumar Ganesan <kg...@yahoo.com>]

Hi

I have 138 documents each of 292 KB.
<ROWLIST> <ROW></ROW> </ROWLIST>
I have 1000 nodes <ROW> </ROW> in each document.
Now i need to do some processing in all the rows.
The xpath query "/ROWLIST" doesn't return anything, this could be because of
the numbers involved 138*100;
Is there any method that would return all the document keys in the database,
so that we can access one document at a time?


Thanks
Satish