Add spec_id back to data file
#63
Conversation
There was a problem hiding this comment.
This is a great start, but I think we need to inherit the spec from the manifest.
I think we can generalize the _inherit_sequence_number by renaming it to something like _inherit_from_manifest, and add a line to copy the spec-id from the manifest:
entry.data_file.spec_id = manifest.partition_spec_id
pyiceberg/manifest.py
Outdated
| @@ -182,6 +182,7 @@ def __repr__(self) -> str: | |||
| doc="Splittable offsets", | |||
| ), | |||
| NestedField(field_id=140, name="sort_order_id", field_type=IntegerType(), required=False, doc="Sort order ID"), | |||
| NestedField(field_id=141, name="spec_id", field_type=IntegerType(), required=False, doc="Partition spec ID"), | |||
There was a problem hiding this comment.
I don't think we want to add the fields here since the spec_id isn't read/written.
|
@puchengy can you fix the CI? We need to make this part of 0.5.1 since the |
|
@Fokko Will do that. |
|
|
|
@Fokko thanks I will address that today or tomorrow. If this becomes a blocker of the release and we need get it done quick, feel free to take it over. |
|
@puchengy let me give it a try |
|
@puchengy I think this fixes it: puchengy#1 |
puchengy
changed the title
|
@Fokko Thanks, adopted your suggestion and tried a local integration test and it passed. |
|
Thanks! It looks like some changes on the request side broke the CI: #69 |
puchengy
force-pushed
the
spec-id-data-file
branch
from
62edada
to
d23e107
Compare
Fokko
changed the title
spec_id back to data file
|
This is great, thanks @puchengy for fixing this 👍 |
Spec id was removed in https://github.com/apache/iceberg-python/pull/40/files to wait for the decision of apache/iceberg#8730. This PR adds it back and is needed to unblock issue #46