AWS S3 agent changelog
Release notes
2.3.2.0
Released: November 10, 2026
-
Parquet: an unconstrained
NUMERICvalue is written with its correct value, instead of at the wrong scale of itsDECIMAL(38,18)column, where0.1was read back as1E-18; -
Parquet: a value with more decimals than the scale of its
DECIMALcolumn is rounded half up with one warning per column, instead of stopping the entity; only a value that does not fit the precision still stops it; -
JSON: numbers inside arrays and maps are written exactly, and
NaNandInfinityare written as strings, instead of failing the entity; -
The files the agent writes can be queried in Query Studio and Query Forge through the agent’s own connection;
2.3.0.1
Released: October 15, 2026
-
CSV files: array and map values are written as JSON text;
-
The agent hands its pooled database back when its pipeline is deleted or detached and when its initialization fails. Before, the next pipeline writing to the same database could fail in a loop with
NoClassDefFoundError;