aloneguid / parquet-dotnet

Fully managed Apache Parquet implementation
https://aloneguid.github.io/parquet-dotnet/
MIT License
600 stars 151 forks source link

[WIP] Untyped dictionary serializer #439

Closed aloneguid closed 9 months ago

aloneguid commented 9 months ago

Use single codebase from FSC/FAC to understand untyped dictionaries. This will deprecate row API when absolutely stable.

Motivation

  1. Single codebase for class serializer and untyped dictionary serializer.
  2. De-serialization produces JSON-like structures in memory. These can be written back to JSON file as is.
  3. Row API is an old legacy that is somewhat buggy and very hard to evolve and fix.

Release plan

This will be released as an addition to the current class serializer in v4. Row API methods will be marked as deprecated and eventually removed in v5, making it a primary dynamic API.

Todo