kskalski commented on issue #4814:
URL: https://github.com/apache/arrow-rs/issues/4814#issuecomment-1718202641
I think `DataType::Null` is a correct data type in this situation and
compatible with other inference behavior, e.g.
```rust
const DATA: &str = r#"{"a": 1, "b": "str", "c": null, "d": []}"#;
```
will create a schema with `d` field, which has a type `List(Field { name:
"item", data_type: Null, nullable: true, dict_id: 0, dict_is_ordered: false,
metadata: {} })`
Creating data frames with such schema is a bit different topic, I guess not
all output formats support null columns, however there is
https://docs.rs/arrow/latest/arrow/array/struct.NullArray.html so:
* up to a point it could still be used to create / process data
* if user wants to use schema for writing to format not allowing nulls (e.g.
parquet seems to be one, though "Parquet Logical Type Definitions" mention
column of `UNKNOWN` type, which relate to arrow null type), they could modify
schema to assign some other datatype with `nullable` option and this should
work fine when parsing data in strict mode
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
