PERF: Improve perf initalizing DataFrame with a range

[topper-123]

Performance improvement when initalizing a DataFrame with a range:

>>> %timeit pd.DataFrame(range(1_000_000)) 
347 ms ± 4.46 ms per loop  # master
5.98 ms ± 61.9 µs per loop  # this PR
# other initialization types don't seem to be affected by this problem:
>>> %timeit pd.Series(range(1_000_000))
5.75 ms ± 27.6 µs per loop  # master and this PR
>>> %t pd.DataFrame({"a": range(1_000_000)})
11.2 ms ± 119 µs per loop  # master and this PR

[WillAyd]

lgtm

[WillAyd]

Can you check we already have a test for this construction?

[jbrockmendel]

nice. does this need/merit an asv?

[topper-123]

Can you check we already have a test for this construction?

It's tested in pandas\tests\frame\test_constructors.py line 1079 and used in a few more places.

[topper-123]

nice. does this need/merit an asv?

IMO that would not be really important, though I'm not against it either. This pattern is probably mostly used interactively, in ipython and jupyter etc. and little otherwise.

[jschendel]

It seems like some care is needed here in respect to dtypes. Specifically if the range contains values only supported by uint64, or values only supported by Python integers.

For example, the following works on master:

In [2]: pd.DataFrame(range(2**63, 2**63 + 4))
Out[2]: 
                     0
0  9223372036854775808
1  9223372036854775809
2  9223372036854775810
3  9223372036854775811

In [3]: _.dtypes
Out[3]: 
0    uint64
dtype: object

In [4]: pd.DataFrame(range(2**73, 2**73 + 4))
Out[4]: 
                        0
0  9444732965739290427392
1  9444732965739290427393
2  9444732965739290427394
3  9444732965739290427395

In [5]: _.dtypes
Out[5]: 
0    object
dtype: object

But both fail with the changes in this PR:

In [2]: pd.DataFrame(range(2**63, 2**63 + 4))
---------------------------------------------------------------------------
OverflowError: Python int too large to convert to C long

In [3]: pd.DataFrame(range(2**73, 2**73 + 4))
---------------------------------------------------------------------------
OverflowError: Python int too large to convert to C long

Admittedly, this is a bit of a corner case. It also looks like the issue isn't limited to the PR, as the Series equivalent of the above fails on master.

[jschendel]

xref #30173 for the failing Series example on master; I'd expect that both issues could be solved in a similar way.

[jreback]

bugs can certainly be addressed in a followup, but can you add an asv ? I think we have a number of construction asv's already

[topper-123]

I've added a ASV. I can take on #30173 in a followup for both Series and Dataframes.

[jreback]

lgtm. merge on green.