First, the features of the statistic were divided into an ambulant and
into an inpatient part, while the total number of medicated patients
was determined separately. This has been done, because the outpatient
medications could be checked more easily, since the total number of
ambulant attendances
is the only information determined.
The subdivision of stationary medications and of the number of days in hospital has been ignored in the beginning. The correct split up was inspected after it became likely that its total is correct. Because the split was done by means of an integer value in the field reason (grund) in the table stat1, testing this subdivision was comparatively simple.
Due to the fact that all features of this statistic
basically only count entries in certain fields of the database,
the main test of it consisted of a volume test. In
other words, the program got a huge amount of data, but no signs of a possible
bug could be found.