Cory Grinstead
Cory Grinstead
I agree. I am actually in the process of converting the pattern to match ECS, and it is quite time consuming. Would be nice if it matched the schema with...
not super familiar with the Python ecosystem, but JS has a [linting rule](https://github.com/typescript-eslint/typescript-eslint/blob/main/packages/eslint-plugin/docs/rules/member-ordering.md) that can enforce this. I wonder if there is a python equivalent.
> Shall we rename json_lines to ndjson? i think both terms are interchangeable, just a matter of preference i guess.
I think those two operations are definitely valuable. I agree that they are not 'technically' lazy operations. That was my initial reasoning behind the `de/serialize` naming convention. It made it...
the lazyframe json methods provide a functionality that doesnt exist in eager. It allows us to instantiate an in memory lazyframe from an a plan, instead of an eager dataset....
This is a result of using hashmaps in the `infer_schema` function within `frame::row`. Hashmaps don't maintain order, I'll work on creating an indexmap equivalent that will maintain the column order....
> One alternative to consider for the above function names is using `load` instead of `extract` => `json_load` and `json_path_load`. This is more consistent with the `json.loads` function in Python....
I'm wondering, what is the benefit of exposing the `json_infer` method? could we not do something similar to the read_* implementations. & take a schema or an infer_count perhaps a...
in spark, there is a notion of a `Dataset` which is a strongly typed version of a dataframe. They are built using scala case classes, so that should be translatable...