Drag and drop open of parquet file

[ClusteredOcelot]

What inspired this wish list request? CSV files are great until they aren't. When they get big they are slow to open, they take up a ton of space on hard drives and the usefulness of having a human readable text file is diminished. Currently parquet files can be opened in jmp 18 by installing some python packages and then using an jsl script that calls python functions (https://community.jmp.com/t5/JMP-Add-Ins/Apache-Parquet-file-importer/ta-p/637823). This is a nice stop gap but having drag and drop opening functionality would be significantly better.

What is the improvement you would like to see? Drag and drop opening of parquet files.

Why is this idea important? CSVs are great for smaller data, it would good to have a solution for larger datasets. 

[Paul_Nelson]

Not planned at this time.  

We believe that the functionality provided by running Python and the PyArrow package serve the need at this time.  JMP 19 brings support for the Portable Dataframe Protocol.  This allows easy conversion from packages that support the dataframe interface (parquet, pandas, ibis, polars,...).  It's basically one line of code to go from a supported dataframe to a  jmp.DataTable or jmp.DataTable to dataframe. At present we are intentionally limiting the packages that JMP ships with, to the Python standard library. Supporting native drag and drop of Parquet files would require JMP to ship with a dependency on PyArrow or a dependency on the C++ Arrow source code.  

There is a large list of things that we want to do to improve the Python Integration's capabilities that will benefit all users, not just those that need a specific file format.  The design philosophy of the Python Integration is to empower the user to build on the tools we supply in JMP, to take JMP further, and customize it to your needs.

Certainly, if there is enough call for parquet to be a native JMP file format, this will be revisited.