Native data iteration from python

[devinrsmith]

I'd like a native way to iterate over Table data from python. I think it can be efficient and batching behind the scenes. Feature wise, I think it needs to ability to iterate all the columns implicitly, or just a subset (without the need to create a view).

# iterate over all columns
for x, y, z in my_table:
  ...
  
# iterate over some columns
for foo, bar in my_table.iter(["Foo", "Bar"]):
   ...

Maybe the interface would need to be context manager based though? If so, then maybe something like:

with my_table.data() as data:
  for x, y, z in data:
    ...

with my_table.data(["Foo", "Bar"]) as data:
  for foo, bar in data:
    ...

I'm not picky about the naming.

As a workaround, I can currently do something like this:

for x, y, z in to_numpy(my_table):
    ...

for foo, bar in to_numpy(my_table, ["Foo", "Bar"]):
    ...

but this is an all-at-once approach that is only to be used with "small" tables.

[devinrsmith]

Potential code in deephaven.table_listener._changes_to_numpy

[chipkent]

How would this work with ticking tables? The iterator could change under you?

[jjbrosnan]

It would be pretty awesome to enable the use of enumerate with something like this as well.