Well, the article title already shows that it is about a rather tricky but maybe not too seldom use case. Assume you have used the medini profiling to customize you data and lets say you use scripting to traverse and accumulate data and return them as "comma separated lists" at the elements. For example in an FTA you are searching for input events that match a certain criteria. Finally you want to see the accumulated data not just in a text cell but in a nice table, so individual cells per item in the CSV - which is challenging with BIRT.
To make it simple I prepared an example for illustration. In that example I have scripted two attributes. The first simply takes the Event name and splits it into individual terms. The second attribute does the same but returns the number of characters in each term.
As you can see, the information is available as an attribute to each Event object and as "comma separate string".
Now you want to put that into a nice report and the challenge is to have rows for each Event in a table but also rows for each entry in the CSV text. As you know, BIRT list and table expansion works solely with data sets that have rows per element so the tricky part is: how to create a data set that extracts the data from the CSV attribute?
There are for sure many potential solutions and I just want to show here one solution that is feasible. I used a normal OCL DS and data set for the FTA events and used a scripted data set to extract comma separated values. The data set (I named it CSV) has just two attributes, "index" and "value" and an "input" parameter. That parameter can be used to pass an arbitrary CSV string into the data set:
We can now create a table that outputs all events (e.g. ID, Name, Kind) and inside each Event row another table that is bound to the CSV data set and parameterized with an attribute from the outer Event object.
Generating the report (for example to Excel) will produce a result similar to this one:
I have attached a medini project that contains (1) the described profiling + some sample data used for the screenshots and (2) the BIRT design file that contains all data sources and sets as well as the shown sample tables. You can use it either to copy stuff to your report design or to play and enhance.
Happy Reporting
Jan
