Hi,


I am developing a db using Berkley DB XML for storing a (fairly large)
document with many nodes. The data is structured in a very simple way
and I am using a Node type container.
Since I am experiencing quite bad performance when running queries I
have started experimenting with indices. The problem is that I don't
see any improvement after creating the index. I have tried many
different types of indices but without luck (queries are still slow).
I can see that the indices have been created ok (by listing them
afterwards) and with the help of the qplan command (in the Berkeley db
xml shell) I can see that the query will use the index.

Just for testing I made up an even simpler structure,looking something
like...

<Root>
<A><B>1</B></A>
<A><B>2</B></A>
<A><B>3</B></A>
<A><B>4</B></A>
<A><B>5</B></A>

Hi Mikael,

It looks like your coming accross a current known issue with the
storage format - materializing a node with a large number of children
is very slow. We are currently looking into ways to solve this problem.

In the mean time, maybe a faster structure for your container would be
to store each "A" element in it's own document in the collection. This
should be a lot faster for you.

John

Thanks :-)
That made a world of difference!

/Mikael