You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
[C++][Dataset] arrow::dataset::Partitioning::Default() can't be used for writing dataset #15256
Is it expected that we can't use arrow::dataset::Partitioning::Default() for writing dataset?
If it's expected, how about removing arrow::dataset::Partitioning::Default() because it's useless?
If it's not expected, how about implementing arrow::dataset::DefaultPartitioning::Format() like the following?
It seems that we want to use arrow::dataset::Partitioning::Default() for no partition write. (All data are written to the base directory.)
I agree. Either Partitioning::Default means "no partitioning" or we get rid of Partitioning::Default and make partitioning optional anywhere it is specified. I think your suggestion (Partitioning::Default means "all in one directory") is the easiest.
Activity
westonpace commentedon Jan 10, 2023
https://issues.apache.org/jira/browse/ARROW-15409 is probably related.
kou commentedon Jan 11, 2023
Thanks! I didn't notice it.
It seems that we want to use
arrow::dataset::Partitioning::Default()for no partition write. (All data are written to the base directory.)westonpace commentedon Jan 11, 2023
I agree. Either
Partitioning::Defaultmeans "no partitioning" or we get rid ofPartitioning::Defaultand make partitioning optional anywhere it is specified. I think your suggestion (Partitioning::Defaultmeans "all in one directory") is the easiest.GH-15256: [C++][Dataset] Add support for writing with Partitioning::D…
apacheGH-15256: [C++][Dataset] Add support for writing with Partition…