Skip to content

[FEA] Dictionary encoding in Parquet writer for nested types #7873

@devavret

Description

@devavret

Nested type support was added to parquet writer in #7230 and #6075 but the data is always plain encoded and not dictionary encoded. This needs significant changes to the dictionary encoding code in parquet writer, which currently hardcodes the hash table size and assumes that the number of rows == number of values.

Metadata

Metadata

Assignees

Labels

cuIOcuIO issuefeature requestNew feature or requestlibcudfAffects libcudf (C++/CUDA) code.

Type

No type

Projects

No projects

Relationships

None yet

Development

No branches or pull requests

Issue actions