Junjun Dong
Junjun Dong
checked_add() not catching underflow/overflow when adding a very large number to a very small number
Hello and first off, thank you for this crate. It seems that checked_add() does not catch addition overflow when adding a very large number to a very small number. For...
# Description `filesystem_check` operation will be rejected if the table is append-only and `files_to_remove` is not empty. It makes sense to have this operation bypass the low-level checking, otherwise append-only...
# Description Currently `DeltaTablePartition::try_from` only supports hive partitioning. If a parquet path has non-hive parititioning, `DeltaTablePartition::try_from` will throw an error. It would be nice to support some non-hive partitioning schemes...
Support parquet files with nanoseconds timestamps when converting a Parquet table to a Delta table
# Description When converting from `ArrowSchema` to `deltalake::Schema`, the `try_from` method will return an error if the arrow schema has a `Timestamp(nanosecond)`. This causes [the `convert_to_delta` function](https://github.com/delta-io/delta-rs/pull/1686) to fail. It...
## Which issue does this PR close? Closes #5436. ## Rationale for this change ## What changes are included in this PR? ## Are these changes tested? ## Are there...
Return a Cow from to_cpu_storage to avoid unnecessary copy. Address one of the issues in https://github.com/huggingface/candle/issues/1699
#4887 introduces IDIV as `BinaryOps.IDIV: lambda x, y: int(x/y) if y != 0 else x*math.inf`. This may return float as a result to integer division. PyTorch errors when handling integer...