Conversation
* introduce polars behind feature flag * add `Dataframe` type * add income dataset which mixes categorical and numerical features * implement records for `polars::DataFrame`
|
|
||
| CsvReader::new(buf) | ||
| .infer_schema(None) | ||
| .has_header(false) |
There was a problem hiding this comment.
Shouldn't this option be given as a function argument has_header: bool?
There was a problem hiding this comment.
Hi, Is not it better to have direct arrow support than providing support for polars. So that any dataframe library based on arrow2 can use the preprocessing, and training methods? I am new to Rust and lifa project my question could be stupid. Thanks in advance for any clarity.
There was a problem hiding this comment.
Hi, Is not it better to have direct arrow support than providing support for polars. So that any dataframe library based on arrow2 can use the preprocessing, and training methods? I am new to Rust and lifa project my question could be stupid. Thanks in advance for any clarity.
But there is so many unique implementations of arrow2 I feel the api should be developed so you can byo arrow2 implementation, this may already an issue, if not I thought I would mention it and maybe I could come up with a poc pr.
4 participants
This is WIP