AsNumpy pythonization is now supported both locally and distributedly. This initial implementation is functional, but not definitive.
AsNumpy works locally with the 6.18 version of ROOT. In order for it to work distributedly some changes to the PyROOT AsNumpy function have to be made.
The changes involve bringing the ndarray class, previously defined inside the function scope, to the global scope. In this way that class would be picklable during the execution of the Spark job.
A few tests have been added, but they will be marked as expected failures until a fix on the PyROOT side will happen. (PR)
The df026_AsNumpyArrays.py tutorial has been added both for the local and the Spark environments.
AsNumpy
pythonization is now supported both locally and distributedly. This initial implementation is functional, but not definitive.AsNumpy
works locally with the 6.18 version of ROOT. In order for it to work distributedly some changes to the PyROOTAsNumpy
function have to be made. The changes involve bringing thendarray
class, previously defined inside the function scope, to the global scope. In this way that class would be picklable during the execution of the Spark job.A few tests have been added, but they will be marked as expected failures until a fix on the PyROOT side will happen. (PR)
The
df026_AsNumpyArrays.py
tutorial has been added both for the local and the Spark environments.