An open-source storage framework that enables building a Lakehouse architecture with compute engines including Spark, PrestoDB, Flink, Trino, and Hive and APIs
Changes the behaviour of generated columns over nested types to allow nullability mismatches like we do for top-level types by switching exact data type equal to using equalsIgnoreNullability instead.
Additionally, improve the error message in the case where the SQL string is identical (which hopefully shouldn't happen anymore after this, but just in case this comes up again later).
How was this patch tested?
Added new tests for various type (mis-) matches.
Does this PR introduce any user-facing changes?
Yes: Generated columns now allow the type of the column definition to differ by nullability from the type of the generating expression, even when the type is a nested type.
Which Delta project/connector is this regarding?
Description
equalsIgnoreNullabilityinstead.How was this patch tested?
Does this PR introduce any user-facing changes?
Yes: Generated columns now allow the type of the column definition to differ by nullability from the type of the generating expression, even when the type is a nested type.