Loading of existing data set without going through string conversions
It is possible to load an existing data set for some post-processing job using InputCppRoot. This algorithm does a conversion to json format as part of the load process. The conversion is necessary as the Inputter can sometimes like to load JobHeader events, sometimes RunHeader events, sometimes Spill events, etc.
Actually, for most analyses we don't really care to load the header and footers, we really only want the Spill data. In this case, a specialist inputter would be faster that did direct load of the C++ data without mucking about with json conversion.
Updated by Rogers, Chris about 5 years ago
- Status changed from Open to Closed
- % Done changed from 0 to 100
Closed in r1017
Note that in the tests, I had to do
maus_cpp.converter.del_data_repr(event) at e.g. line 109 of
src/input/InputCppRootData/test_input_cpp_root_data.py. Not sure how that will work in the single_threaded.py... worth being aware of a potential issue.