[read_root] WIP: Add support for default values in empty arrays

scikit-hep / root_pandas

A Python module for conveniently loading/saving ROOT files as pandas DataFrames

MIT License

109 stars 35 forks source link

[read_root] WIP: Add support for default values in empty arrays #74

Closed pseyfert closed 4 years ago

pseyfert commented 6 years ago

root_numpy supports branch specifications like

branches = ['event_ID', ('jet_E', -99)]

To set default values in empty arrays (and another tuple element to spell out the size of the array).

This patch should enable the user to pass these through read_root:

columns = ['event_ID', ('noexpand:jet_E[0]',-99), ('jet_p?',-88)]

WIP:

[ ] want to see what travis has to say
[ ] no test for the tuple behaviour in place

coveralls commented 6 years ago

Coverage decreased (-2.07%) to 81.757% when pulling 054a704e0d918e9cb568cbc58a9dd3b567eddf96 on pseyfert:tuples_for_branchspec into bf6b624e2e647e7f7efd8823d7515d896afa17ed on scikit-hep:master.

eduardo-rodrigues commented 4 years ago

As for the other PR, shall this not be closed at this stage given what uproot and awkward-array offer? Seems fine to be. Let us know @pseyfert.

pseyfert commented 4 years ago

closing all my cern related PRs as won't-pursue-anymore w/o checking status