Azure Data Lake Storage Gen2 Agent Changelog

Added additional logs around files eviction;

Fixed a regression where parquet files are split again per operation instead of writing everything to the same file;

GSSD-587: Added support for Parquet file with truncate operation;

GSSD-571: Optinally create sidecar JSON compation file;

New lib for parquet file creation made entirely by MOLO17 that comes with 100x performance boost compared to Apache Parquet Java lib;

GSSD-562: Source fields order is now preserved when using Parquet file storage;

GSSD-545: Parquet file can now be stored within a different path strategy (relative timestamp paths, flat paths), see for more details;

GSSD-448: Parquet files made out of snapshot' records now also contain transaction-related metadata;

Updated vendor libraries.

GSSD-448: Applied changes to the Parquet file storage to match the new behavior of the target agent, see Parquet files support for the updated behavior.

Improved error handling;

Fixed Windows compatibility issue;

Perfomance improvement when using Parquet file storage;

GSSD-381: Fixed an issue with SmallInt data types when using Parquet file storage;

GSSD-382: Avoid losing data when parquet files is created involving multiple different change operations simultaneously;

Added support for Parquet file storage;

N/A (initial release)