For Ben to use, to push Arrow tables that he first pulls from Kafka then wants to write into Skyhook.
The Arrow tables are created by pyarrow, and contain some number of complete columns from the events tree. These tables are currently no larger than 14MB.
The write API could be something like:
skyhook.write(datasetname, data, dataformat=arrow)
where data format is one of our supported formats: csv, json, nanoaod, root, etc.
For Ben to use, to push Arrow tables that he first pulls from Kafka then wants to write into Skyhook.
The Arrow tables are created by pyarrow, and contain some number of complete columns from the events tree. These tables are currently no larger than 14MB.
The write API could be something like:
skyhook.write(datasetname, data, dataformat=arrow)where data format is one of our supported formats: csv, json, nanoaod, root, etc.