PERF: get_loc

[jbrockmendel]

Fixes a perf regression from 1.2.x

from asv_bench.benchmarks.indexing import *
setup()
self = NumericSeriesIndexing()
self.setup(UInt64Index, "unique_monotonic_inc")

%timeit self.time_getitem_scalar(UInt64Index, "unique_monotonic_inc")
6.34 µs ± 257 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)  # <- PR
46.5 µs ± 724 ns per loop (mean ± std. dev. of 7 runs, 10000 loops each)  # <- 1.2.0
1.43 ms ± 36.9 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)  # <- master

%timeit self.time_loc_scalar(UInt64Index, "unique_monotonic_inc")
12.6 µs ± 163 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)  # <- PR
53.9 µs ± 1.1 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)  # <- 1.2.0
1.38 ms ± 5.33 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)  # <- master


%timeit self.time_loc_slice(UInt64Index, "unique_monotonic_inc")
32.5 µs ± 2.63 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)  # <- PR
75.6 µs ± 1.22 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)  # <- 1.2.0
1.41 ms ± 10.4 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)  # <- master