PERF: op(frame, series) when series is not EA

[jbrockmendel]

We have the same optimization in place for the flex ops.

ipython results for the asv this adds:

In [3]: arr = np.arange(10 ** 6).reshape(100, -1)                                                                                                                                                                   
In [4]: df = pd.DataFrame(arr)                                                                                                                                                                                         
In [5]: df["C"] = 1.0                                                                                                                                                                                               
In [6]: row = df.iloc[0]                                                                                                                                                                                            

In [11]: %timeit df + row                                                                                                                                                                                           
1.92 s ± 71.1 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)   # <-- master
1 ms ± 19.2 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)  # <-- PR

[jbrockmendel]

@WillAyd any guesses why the benchmark check would be exiting at 50%?

[jbrockmendel]

hmm when i run asv dev locally it stops at 50% but with exit code 0. any guesses @datapythonista ?

[datapythonista]

hmm when i run asv dev locally it stops at 50% but with exit code 0. any guesses @datapythonista ?

I tried asv run locally with master, and it worked well until the end. Is it just asv dev, or just this PR that is stopping half way? I don't see why this should happen.

[jreback]

lgtm. can you add a whatsnew note & an asv that covers this.

[datapythonista]

Tried asv dev locally, and also stopped at 50%. Is there an issue for it? I see that last release of asv was a year ago, so I guess it's something in our side.

[jbrockmendel]

Looks like the asv issue is actually that this causes an unrelated benchmark to fail. Am troubleshooting now and will add an appropriate test.

[jbrockmendel]

lgtm. can you add a whatsnew note & an asv that covers this.

done