armitage420 commented on code in PR #5779:
URL: https://github.com/apache/hive/pull/5779#discussion_r2349557886
##########
serde/src/java/org/apache/hadoop/hive/serde2/avro/AvroDeserializer.java:
##########
@@ -220,12 +223,60 @@ private List<Object> workerBase(List<Object> objectRow,
Schema fileSchema, List<
Object datum = record.get(columnName);
Schema datumSchema = record.getSchema().getField(columnName).schema();
Schema.Field field =
AvroSerdeUtils.isNullableType(fileSchema)?AvroSerdeUtils.getOtherTypeFromNullableType(fileSchema).getField(columnName):fileSchema.getField(columnName);
+ if (field != null) {
+ // This single call handles all cases: direct, union, map, and array.
+ logicalType = findNestedLogicalType(field, field.schema());
+ }
objectRow.add(worker(datum, field == null ? null : field.schema(),
datumSchema, columnType));
}
return objectRow;
}
+ /**
+ * Recursively traverses a schema to find the first LogicalType.
+ * This handles direct logical types, and logical types nested within
+ * UNIONS, MAPS, or ARRAYS.
+ *
+ * @param field
+ * @param schema The schema to inspect.
+ * @return The found LogicalType, or null if none is found.
+ */
+ private LogicalType findNestedLogicalType(Schema.Field field, Schema schema)
{
+ // Base Case 1: The schema is null, so we can't proceed.
+ if (schema == null) {
+ return null;
+ }
+
+ // Base Case 2: The schema itself has the logical type. We found it.
+ if (schema.getLogicalType() != null) {
+ return schema.getLogicalType();
+ }
+
+ // Recursive Step: The logical type is deeper. Check the container type.
+ switch (schema.getType()) {
+ case UNION:
+ // Find the first non-null branch and search within it.
+ return schema.getTypes().stream()
+ .filter(s -> s.getType() != Schema.Type.NULL)
+ .map(s -> findNestedLogicalType(field, s)) // Recurse on the
branch
+ .filter(java.util.Objects::nonNull)
+ .findFirst()
+ .orElse(null);
+ case MAP:
+ // Search within the map's value schema.
+ return findNestedLogicalType(field, schema.getValueType());
+ case ARRAY:
+ // Search within the array's element schema.
+ return findNestedLogicalType(field, schema.getElementType());
+ default:
+ // This type (e.g., STRING, INT) doesn't contain a nested schema.
+ return field.getProp("logicalType") != null
+ ? new LogicalType(field.getProp("logicalType"))
+ : null;
+ }
Review Comment:
The tests for Union type schema is present in avro_hybrid_mixed_timestamp.q
file:
<img width="1864" height="974" alt="image"
src="https://github.com/user-attachments/assets/e54d6640-c317-4ab5-9367-249fc03819a0";
/>
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
