Itamar Ravid
Itamar Ravid
Chunk#foreach is Chunk#mapM :-) might make sense to rename or alias.
The security manager is slated for removal: https://openjdk.java.net/jeps/411 Maybe we could instead add a method to our zio.system.System service for abruptly shutting down a ZIO-based app that would bypass the...
Thanks for the explanation @raboof! Agreed that this is not a false positive; sounds very important for Java/Kotlin interop usecases.
Also refs #164
@OlivierBlanvillain have another look; this approach should allow to keep the column reference in a value, for example, and then have the actual TypedColumn constructed when using select.
Ah, but this approach doesn't support expressions like `val c = functions.col('a) + 1`. Tradeoffs, tradeoffs...
What's the reasoning behind carrying around the dataset type in the typed columns?
That's true, but how about moving the column existence evidence to the methods on `TypedDataset`? The reasoning is that a `TypedColumn` can freely exist outside the context of a typed...
It is isomorphic up to the requirement of having `ds` at hand ;) I will open a separate PR to experiment with moving the evidence to `ds.select` et al and...
Thanks @OlivierBlanvillain, could you provide some historical perspective on the second approach? It seems similar to the old `TypedDataFrame` description in the docs but I am not aware of how...