Details
-
Bug
-
Status: Closed
-
Major
-
Resolution: Fixed
-
1.1.0, 3.0.0
-
None
-
Reviewed
Description
The following methods in AvroDeserializer keeps regenerating TypeInfo objects for every nullable field in a row.
This is happening in the following methods.
private Object deserializeNullableUnion(Object datum, Schema fileSchema, Schema recordSchema) throws AvroSerdeException { // elided line 312: return worker(datum, fileSchema, newRecordSchema, SchemaToTypeInfo.generateTypeInfo(newRecordSchema, null)); } .. private Object deserializeSingleItemNullableUnion(Object datum, Schema Schema recordSchema) // elided line 357: return worker(datum, currentFileSchema, schema, SchemaToTypeInfo.generateTypeInfo(schema, null));
This is really bad in terms of performance. I'm not sure why didn't we use the TypeInfo we already have instead of generating again for each nullable field. If you look at the worker method which calls the method deserializeNullableUnion the typeInfo corresponding to the nullable field column is already determined.
Moreover the cache in SchemaToTypeInfo class does not help in nullable Avro records case as checking if an Avro record schema object already exists in the cache requires traversing all the fields in the record schema.
I've attached profiling snapshot which shows maximum time is being spent in the cache.
One way of fixing this IMO might be to make use of the column TypeInfo which is already passed in the worker method.
Attachments
Attachments
Issue Links
- relates to
-
HIVE-18410 [Performance][Avro] Reading flat Avro tables is very expensive in Hive
-
- Closed
-
- links to