Github user KanakaKumar commented on a diff in the pull request:
https://github.com/apache/carbondata/pull/2687#discussion_r216122590
--- Diff:
store/sdk/src/main/java/org/apache/carbondata/sdk/file/AvroCarbonWriter.java ---
@@ -198,14 +199,35 @@ private Object avroFieldToObject(Schema.Field
avroField, Object fieldValue) {
if (unionField.getType().equals(Schema.Type.NULL)) {
continue;
}
- if (checkFieldValueType(unionField.getType(), fieldValue)) {
+ // Union may not contain more than one schema with the same type,
+ // except for the named types record,fixed and enum
+ // hence check for schema also in case of union of multiple
record or enum or fixed type
+ if (checkFieldValueType(unionField.getType(), fieldValue,
unionField)) {
values[j] = avroFieldToObjectForUnionType(unionField,
fieldValue, avroField);
break;
}
j++;
}
out = new StructObject(values);
break;
+ case BYTES:
+ // DECIMAL type is defined in Avro as a BYTE type with the
logicalType property
+ // set to "decimal" and a specified precision and scale
+ // As binary type is not supported yet,value will be null
--- End diff --
We should throw unsupported data type exception for bytes. Don't store null.
---
