rdblue commented on code in PR #7392:
URL: https://github.com/apache/iceberg/pull/7392#discussion_r1224541725
##########
core/src/main/java/org/apache/iceberg/avro/AvroWithPartnerByStructureVisitor.java:
##########
@@ -93,14 +94,23 @@ private static <P, T> T visitRecord(
private static <P, T> T visitUnion(
P type, Schema union, AvroWithPartnerByStructureVisitor<P, T> visitor) {
List<Schema> types = union.getTypes();
- Preconditions.checkArgument(
- AvroSchemaUtil.isOptionSchema(union), "Cannot visit non-option union:
%s", union);
List<T> options = Lists.newArrayListWithExpectedSize(types.size());
- for (Schema branch : types) {
- if (branch.getType() == Schema.Type.NULL) {
- options.add(visit(visitor.nullType(), branch, visitor));
- } else {
- options.add(visit(type, branch, visitor));
+ if (AvroSchemaUtil.isOptionSchema(union)) {
+ for (Schema branch : types) {
+ if (branch.getType() == Schema.Type.NULL) {
+ options.add(visit(visitor.nullType(), branch, visitor));
+ } else {
+ options.add(visit(type, branch, visitor));
+ }
+ }
+ } else {
+ List<Schema> nonNullTypes =
+ types.stream().filter(t -> t.getType() !=
Schema.Type.NULL).collect(Collectors.toList());
+ for (int i = 0; i < nonNullTypes.size(); i++) {
+ // In the case of complex union, the corresponding "type" is a struct.
Non-null type i in
+ // the union maps to struct filed i + 1 because the first struct field
is the "tag".
+ options.add(
+ visit(visitor.fieldNameAndType(type, i + 1).second(),
nonNullTypes.get(i), visitor));
Review Comment:
This looks fine, except for the way null is handled.
If the visitor implementation chooses to ignore the null type, that's fine.
But it I think it should still be visited. Since there is no matching struct
field, you can pass `null` for the field.
--
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]
