Scripted Datasource Peformance Issues

bigtruck

Hi,
I am having some performance issues with scripted datasets that perhaps someone here will have a solution for.

I bring data in via a scripted dataset, and the result set is about 14,000 rows, with about 15 columns. From what I can tell, this means that the fetch() script for the dataset is called 14,000 times. After timing this, I found that it takes about 1.5 minutes for BIRT to iterate through the 14,000 rows. It seems to me that it should be much faster than that. The data is coming from a Java Arraylist object.

Even if I run the fetch() script 14,000 times without bringing any data in, the performance is still the same.

I'm wondering if there might be a couple solutions to get around this...

1. I have some setting which needs to be corrected.
2. I bring in the entire 14,000 row Java object all at once in the open() script, and somehow manually build a dataset from that by creating a loop in the open() script (instead of calling the fetch() script 14,000 times).

Thoughts?

mwilliams

Hi bigtruck,

I don't believe there are any settings to be able to speed up a scripted dataSource. Can you attach your report design and java file if you are using one so that I can see exactly what you're doing?