Module spark_compat 

Spark-compatible UDF implementations.

Functions that match Apache Spark / Databricks behaviour so existing Spark SQL transformations run unchanged inside the embedded engine.

Functions

make_millis_to_ts_ntz_udf

millis_to_ts_ntz(millis) — INT64 milliseconds since epoch → timezone-naive microsecond timestamp (Timestamp(µs, None)).

make_millis_to_ts_udf

millis_to_ts(millis) — INT64 milliseconds since epoch → timezone-aware (UTC) microsecond timestamp.

make_sha2_udf

sha2(expr, bit_length) — Databricks/Spark-compatible SHA-2 returning a lowercase HEX STRING (Utf8), NOT raw bytes.

make_to_timestamp_ms_ntz_udf

to_timestamp_ms_ntz(millis) — INT64 milliseconds since epoch → timezone-naive microsecond timestamp (Timestamp(µs, None)), the tz-naive counterpart of make_to_timestamp_ms_udf.

make_to_timestamp_ms_udf

to_timestamp_ms(millis) — timezone-aware (UTC) alias of make_millis_to_ts_udf under a Spark-familiar name.