No way of manipulating data tables in memory

[as2875]

@lwinfree, @sje30

When converting to Frictionless from another data format (e.g. HDF5), scripts have to

1. import the package descriptor,
2. make a package with the descriptor,

   datapackage_path = pkg_resources.resource_filename(
   __package__,
   "datapackage.json")
   package = datapackage.Package(
   base_path=base,
   descriptor=datapackage_path)

3. make some CSV files with the right names, and finally
4. call package.save.

It would be useful to skip the stage with writing CSV files to disk. If this functionality exists and I am missing something, please let me know, it would be very useful.

---

Please preserve this line to notify @roll (lead of this repository)

[lwinfree]

hi @roll when you are back next week, can you please look at this?

[roll]

Hi @as2875,

could you please elaborate a little bit? You mean you don't want package.save to save the data, only the descriptor?

[as2875]

Hi @roll. I mean that I don't want to write CSV files to the disk, just the final zipped data package. Say I have some tables stored in Python data structures in memory. Rather than write the tables to CSV files, then call package.save, I would like to create some Resource objects, point them to the data structures, and then call package.save.

The example is from converting multiple HDF5 files to Frictionless data packages. At the moment, I have to create a package, read in the HDF5 datasets, store them as 2-D lists, write the contents of the lists to CSV files, call package.save, then delete the original CSV files. I want to skip the operations involving CSV files.

[roll]

Thanks. I think it's not possible at the moment. I've marked it as feature request

[as2875]

Thanks @roll. This would make data conversion pipelines and parallel processing a lot simpler.

[roll]

@as2875 BTW there is dataflows - https://github.com/datahq/dataflows, I'm wondering you can achieve this goal using a flow

[as2875]

Thanks for the suggestion @roll. dataflows looks promising.

[roll]

MERGED into https://github.com/frictionlessdata/frictionless-py/issues/439

More info about Frictionless Framework