The GitHub Actions job "Docs 404 check (beta)" on flink.git/FLINK-41027 has 
failed.
Run started by GitHub user kumarpritam863 (triggered by kumarpritam863).

Head commit for run:
c2d3c9acae26ae9642c7550e5e7d1222baff26ef / Pritam Kumar Mishra 
<[email protected]>
[FLINK-41027][format/avro] Add an option to map time-millis and time-micros to 
INT and BIGINT

An Avro time-millis becomes TIME(3) and a time-micros TIME(6) in Flink. A TIME
holds milliseconds only, so the microseconds of a time-micros do not survive
it, and a TIME(6) has no Avro schema at all: AvroSchemaConverter.convertToSchema
refuses a precision above 3. A time-micros field in the 'schema' of 
avro-confluent
could therefore not be used at all: a TIME(6) column failed before the provided
schema was looked at, and a BIGINT column did not match it.

Add a flag that keeps the Avro representation, time-millis as an int and
time-micros as a long:

- AvroSchemaConverter.convertToDataType and convertToTypeInfo take
  timeAsNumber and then return INT / BIGINT (Types.INT / Types.LONG) for the
  two logical types. The existing overloads keep the current mapping.
- AvroToVariantDataConverters.createVariantConverter takes timeAsNumber. With
  it time-millis stays the Avro int instead of being converted to microseconds
  of day and stored as a long.
- avro-confluent gets 'time-mapping.as-number' (default false), which applies
  to the table schema check against the provided 'schema'. It is documented in
  both languages. debezium-avro-confluent is not changed.

RegistryAvroFormatFactory also built the Avro schema of the table with
Optional.orElse, which runs even when a schema is provided, so a TIME(6) column
failed there before the provided schema was looked at. It now uses orElseGet.

The default is unchanged apart from that.

Generated-by: Claude Code

Report URL: https://github.com/apache/flink/actions/runs/39247265459

With regards,
GitHub Actions via GitBox

Reply via email to