PERF: RangeIndex.is_monotonic_inc/dec

[sinhrks]

• tests added / passed
• passes git diff upstream/master | flake8 --diff
• whatsnew entry

as the same as is_unique, it should be monotonic by definition

%timeit -n 1 -r 1 pd.RangeIndex(0, 1000000).is_monotonic_increasing

# current master
#1 loop, best of 1: 17.2 ms per loop

# this PR
#1 loop, best of 1: 45.1 µs per loop

[codecov-io]

Current coverage is 84.57% (diff: 100%)

No coverage report found for master at b60e42b.

Powered by Codecov. Last update b60e42b...8e25563

[jorisvandenbossche]

Maybe add a test for the empty case as well?

In [23]: pd.RangeIndex().is_monotonic_increasing
Out[23]: True

In [24]: pd.RangeIndex().is_monotonic_decreasing
Out[24]: True

[jorisvandenbossche]

Although this should maybe rather raise an error (@jreback)

Other index types to raise without arguments, and range() does as well.

[sinhrks]

@jorisvandenbossche empty case is coverd by (1, 1).

pd.RangeIndex(1, 1).values
# array([], dtype=int64)

Let me issue a separate PR to prohibit empty construction.

[jorisvandenbossche]

Ah yes, assumed that RangeIndex() does not has a step yet, but that is of course wrong :-)

[jreback]

yeah I don't think RangeIndex() should be permitted, though I seem to remember needing it for unpickling or somesuch.......

[jreback]

let's make a new issue about the empty RangeIndex()