aggregate_logger: fix bug with timestamp at beginning of the new year

@schaefed

The problem was, that a a aggregated year should start with the timestamp 00:30 when aggregated to 30 minutes and with the 01.01. when aggregated to a whole day.
The new solution seems to be a bit hacky.
In the writeData function a freq is needed (I renamed is to change_freq) for the iterDataFreq(data, change_freq) function in dacces.py Until now change_freq=y, that isn't even listed in freqs dictionary in the iterDataFreq function in dacces.py , but it works only for aggregation for 1d, not for the 30min aggregation.
For the 30min it works with change_freq=yearly.
Maybe you have an idea how to solve this better.

The doall function has always failed because the CR1_KS1_Erms_2012_03_14_0010.level1.csv file is empty. This should be deleted first.

added 1 commit

• dc5e8d1c - do_logger_aggregation: changes for the timestamp bug and change in dacces for skipping empty files

Compare with previous version

resolved all threads

mentioned in commit 6e1739a6

merged

Merged and deployed!