Conversation
|
Added the Numpy ReadArrays Added the writer access from table class |
solatis
left a comment
solatis
left a comment
There was a problem hiding this comment.
This is not the right implementation. You should have taken the bulk reader from the pandas API (with streaming support that yields individual batches, because that's the whole reason the bulk reader exists -- streaming over a table rather than ingesting everything in memory).
Then the pandas adapter wraps the numpy adapter, like it does with all other APIs, and we don't need to have two different consumers of the bulk reader. Then it mirrors the approach of the writer and our old column-oriented API. I'm not sure why the bulk reader was implemented in the pandas adapter, that must have been an oversight.
| table: Optional[Union[str, Table]] = None, | ||
| columns: Optional[Sequence[str]] = None, | ||
| ranges: Any = None, | ||
| ) -> Tuple[NDArrayTime, Dict[str, MaskedArrayAny]]: |
| qdb::writer_ptr table::writer() const | ||
| { | ||
| _handle->check_open(); | ||
| return std::make_unique<qdb::writer>(_handle); | ||
| } | ||
|
|
There was a problem hiding this comment.
Why is this necessary? A writer operates on multiple tables
2 participants
No description provided.