paleolimbot commented on code in PR #174:
URL: https://github.com/apache/arrow-nanoarrow/pull/174#discussion_r1151298323
##########
src/nanoarrow/nanoarrow_types.h:
##########
@@ -503,7 +503,8 @@ struct ArrowArrayView {
///
/// If storage_type is a union type, a 256-byte ArrowMalloc()ed buffer
/// such that child_index == union_type_id_map[type_id] and
- /// type_id == union_type_id_map[128 + child_index]
+ /// type_id == union_type_id_map[128 + child_index]. This value may be
+ /// NULL in the case where child_id == type_id.
Review Comment:
This is an implementation detail of the `ArrowArrayView`...in the spec the
type ids are always there. Arguably they should always be here, too, but
there's at least one test that exploits that behaviour that I discovered so I
figured I should probably document it. I think the reason it ended up that way
was because you can `ArrowArrayViewInitFromType()` and
`ArrowArrayViewAllocateChildren()` and letting the type id map stay NULL saves
some special casing of unions there.
--
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]
