Data

What's the fun of a reporting tool without data ? Pretty much no fun at all, so I decided to find what the internet had to offer.

In the interest of learning More New Stuff, I installed MySQL and searched for sample databases. Sakila looked hopeful, but appears to be algorithmically generated film details... which is nice for seeing what MySQL does, but is somewhat lacking as far as fun details is concerned.

So, a few sample reports were enough to make me decide that Sakila wasn't the data to learn BIRT with. Meatier data would be much more appealing.

However, it did make it apparent that BIRT appears to be a lot more a *Reporting* tool than a "Business Intelligence" tool (at least in the Business Objects sense). The user needs to be able to build the queries to make sensible reports - although BIRT does allow separate datasets to be joined together in Eclipse.

The  search for more interesting data led to the CIA World Factbook. Downloadable copies from 2000-2010 are available, plus the "current" version. The country comparison data appears in 3 forms - web, print and raw - in a the "rankorder" subfolder of the factbook. Each comparisopn has a mystery 4-digit identifier which links files XXXXrank.html, print_XXXX.html and rawdata_XXXX.txt. Sadly the "raw" data is simply tab-separated text files with no column headings, and missing the minimal metadata included in the web/print versions (principally the year the data value was originally estimated). So, some parsing of the print or web versions may be necessary to decode the identifiers into just what data is provided... plus some data from the print/web versions may be useful too.

However, that's me getting ahead of myself. It is Real Data and it should let me see some of what BIRT can do. BIRT includes the ability to defined Flat File data sets, so click, select file, click, click, click and... nothing useful.The entire file appearing as a single exceedingly long record. Not what was expected at all.