reswqa

Thanks for the PR, but I'm not sure about this. AFAIK, processing time doesn't make much sense in batch mode. In particular, we return a processing time for event time...

We shouldn't be able to infer a shape from `(-1, -1)`, maybe explicitly error instead of panic is enough, right?

> I think one of (x, y) must be "positive"? I'm not sure what 0 is supposed to do: Fair enough, the result of `(0,0)` is weird.

> Possible solutions include selecting the appropriate way to read the data based on the PartitionType or adding a new field, byteOrder, to explicitly inform the server of the endianness...

You can shift the paradigm to `when-then-otherwise` as `set` is a little anti-pattern, and will block optimizations like predicate pushdown... ```python pl_series_datetime = pl.Series("date", [datetime.date(1751, 1, 1), datetime.date(1754, 1, 1)])...

When deciding the schema for `Ternary`, we let the name of `truthy` branch be the name of `Ternary` expression. We should also follow this with `RenameAlias`, so I think it's...

That's the purpose of `name.keep`, right? ```python pl.when(pl.col.c).then(None).otherwise(pl.all()).name.keep() ```

> I assume that it, like everything else, looks for the left-most expression, which means that .name.keep() should be failing as well, just like .name.suffix() now does too. The key...

> until someone runs into an issue What do you mean by `issue`? The test case you have to workaround? I can open a PR(It might be a little late...

> The "root name" (I think) should be what's in the pl.when expression (as of this PR), which is now "literal" Depending on how you define the `root expr` for...