[Python] Add factory function to create a Table from Columns and Schema.

[asfimport]

At the moment it seems to be impossible in Python to add custom metadata to a Table or Column. The closest I've come is to create a list of new Fields (by "appending" metadata to existing Fields), and then creating a new Schema from these Fields using the Schema factory function. But I can't see how to create a new table from the existing Columns and my new Schema, which I understand would be the way to do it in C++?

Essentially, wrappers for the Table's Make(...) functions seem to be missing.

Reporter: Thomas Buhrmann / @buhrmann

_{Note: This issue was originally created as ARROW-2572. Please see the migration documentation for further details.}

[asfimport]

Antoine Pitrou / @pitrou:
Actually, Table.from_arrays allows you to pass columns (instead of arrays) and an optional schema. Since that's not very discoverable, we may want to add a Table.from_columns alias to that method. Also, the Table.from_arrays docstring doesn't mention the schema parameter.

[asfimport]

Antoine Pitrou / @pitrou:
If you want to try out contributing, this would be a good issue to start and submit a PR for.

[asfimport]

Thomas Buhrmann / @buhrmann:
Another (discoverable) workaround was to convert the table to record batches and then use the from_batches function, but yes, the aliases would be much nicer. 

Perhaps next week I may try to set up a dev environment and see whether I can contribute. Thanks for the nudge ;)

[asfimport]

Wes McKinney / @wesm:
Do you want to try contributing a patch?

[asfimport]

Antoine Pitrou / @pitrou:
@jorisvandenbossche

[asfimport]

Joris Van den Bossche / @jorisvandenbossche:
The Table.from_arrays nowadays clearly mentions that it accepts both arrays as columns: https://arrow.apache.org/docs/python/generated/pyarrow.Table.html#pyarrow.Table.from_arrays (and also the schema is now mentioned in the docstring).

Of course, when looking for a method to create a Table from columns, I agree a from_columns is certainly more discoverable. But it might also be API clutter ..

Would we want to make it an exact alias? Or actually a separate function that is more strict (not allowing a list of arrays)? Because for a list of columns, you would not necessarily need to pass the column names (or schema), as is required in from_arrays.

[asfimport]

Wes McKinney / @wesm:
I don't think we need to add an alias. Perhaps add a section in the documentation if there is not one currently

[asfimport]

Wes McKinney / @wesm:
I removed from 0.14.0 milestone but since this is a minor documentation fix please feel free to submit a PR

[asfimport]

Wes McKinney / @wesm:
Now that arrow::Column is gone it would be good to look at this again

[asfimport]

Joris Van den Bossche / @jorisvandenbossche:
Given that Columns no longer exist, I think this can be closed.

If you have defined fields (eg to add custom metadata), you can pass them in a schema to create the table:

field1 = pa.field('field1', pa.int64(), metadata=dict(a="A", b="B")) 
field2 = pa.field('field2', pa.int64(), nullable=False)

table = pa.Table.from_arrays([pa.array([1, 2]), pa.array([3, 4])], schema=pa.schema([field1, field2]))