Is there a plan to support the zstd compression of parquet files?

The parquetread function does not support the zstd compression so I'm forced to write a bunch of ugly code using the python bindings.

Is there any plan to support the zstd compression in any upcomming releases?

[Qianqian]

try my Zmat toolbox (https://github.com/NeuroJSON/zmat/releases/tag/v0.9.9). In v1.0RC1 released in Oct 2023, I added support for zstdencode/zstddecode support to compress/decompress zstd memory buffers.

Precompiled mex files for MATLAB and Octave are embedded in the package so you don't need to build on your own. Aside from zstd, zmat also supports zlib/gzip/lz4/lz4hc/lzma/lzip and the high performance blosc2 meta-compressor (https://blosc.org) for ND-arrays.

Answer 1

[Walter Roberson]

No? Yes? Maybe? Rutabaga ?

That is proprietary information. You will need to contact your sales person and arrange a Non Disclosure Agreement before Mathworks will discuss it with you.

[Rueben Mendelsberg]

Ive asked similar questions and have gotten answers from matlab staff before so thanks but no thanks for your non helpful answer

[Walter Roberson]

Tiime until the appropriate Project Manager (or higher) wanders along and happens to be willing to discuss the matter in public: days to multiple years (e.g., two years for your most recent question involving increment and run in the editor)

Time until you were redirected to the way to official process for dealing with such questions: minutes.