Question about BIRT concepts

grather

I am a bit new to the BIRT world so I apologize if I use incorrect terminology.

I have recently tried to do some basic charting.
Trying to create a line chart.

My raw data looks somthing like this

name rating
joe 3
sam 2
ed 3

I am using a scripted dataset calling some java code.

Instead of returning the raw data. I actually did the processing my self
returning

rating count
3 2
2 1


Was able to easily create the chart and call and run using the api.

But struggled when trying to create the table of data. Another question will be coming

---

So my question is. Does BIRT really like to have the raw data
Rows and columns and then let the report do the aggregation?

Or should I do it on the data source side?

Should I do the counting of the data or should I just send the complete data set to BIRT?

Virgil Dodson

Hi grather,

This really depends on what you need. If you are only going to be displaying in aggregate then it is better to do the rolling up of data in your query especially if you are using a database which are designed to do this quite efficiently. If you will need to display all of the detail rows and the aggregates, then it is better to do the grouping of data within BIRT. There are also several other scenarios where BIRT is better to roll up the data, like if you are using csv, or XML data sources, or combining data sources then processing within BIRT is your only option.