BUG: DatetimeIndex cannot parse string ndarray with dayfirst

[sinhrks]

Closes #5917.

[jreback]

can you run a perf check and post the outliers (i dont' think its a problem), but just checking

[jreback]

@sinhrks can you post a perf check?

[jreback]

@sinhrks need a perf check on this

[sinhrks]

@jreback Sorry to delay. Rebased, and following is vbench results. Looks OK as there is no tests get slower constantly.

# 1st run
-------------------------------------------------------------------------------
Test name                                    | head[ms] | base[ms] |  ratio   |
-------------------------------------------------------------------------------
groupby_pivot_table                          |  43.2317 |  36.2120 |   1.1939 |
frame_reindex_axis1                          | 372.5393 | 306.3263 |   1.2162 |
groupby_transform                            | 405.6187 | 328.8047 |   1.2336 |
groupby_multi_different_numpy_functions      |  30.6609 |  24.5076 |   1.2511 |
groupby_indices                              |  19.0316 |  14.9120 |   1.2763 |
groupby_multi_series_op                      |  33.6157 |  26.0634 |   1.2898 |
groupby_multi_count                          |  18.2153 |  14.0044 |   1.3007 |
groupby_multi_different_functions            |  30.1137 |  22.8820 |   1.3160 |
groupby_sum_booleans                         |   2.6333 |   1.9503 |   1.3502 |
dataframe_reindex                            |   1.1314 |   0.8047 |   1.4060 |
groupby_last                                 |   9.5390 |   6.7446 |   1.4143 |
-------------------------------------------------------------------------------

# 2nd run
-------------------------------------------------------------------------------
Test name                                    | head[ms] | base[ms] |  ratio   |
-------------------------------------------------------------------------------
frame_drop_dup_inplace                       |   6.6934 |   5.7159 |   1.1710 |
reindex_fillna_backfill                      |   1.0877 |   0.8977 |   1.2116 |
frame_constructor_ndarray                    |   0.2096 |   0.1643 |   1.2756 |
ctor_index_array_string                      |   0.0536 |   0.0420 |   1.2784 |
frame_get_numeric_data                       |   0.3580 |   0.2743 |   1.3050 |
series_constructor_ndarray                   |   0.0536 |   0.0386 |   1.3889 |
index_from_series_ctor                       |   0.0663 |   0.0473 |   1.4017 |
-------------------------------------------------------------------------------

[cpcloud]

Are those against latest master?

[sinhrks]

Based on latest master on May 9th. Anything unexpected?

Target [bdb2f8f] : modified
Base   [16a8902] : Merge pull request #7081 from TomAugspurger/fwf-slice

BUG: read_fwf colspec should treat None like slice

[sinhrks]

Rebased to latest and run perf again.

-------------------------------------------------------------------------------
Test name                                    | head[ms] | base[ms] |  ratio   |
-------------------------------------------------------------------------------
reindex_fillna_backfill                      |   1.0796 |   0.9297 |   1.1613 |
timeseries_1min_5min_mean                    |   1.3570 |   1.0761 |   1.2611 |
frame_get_numeric_data                       |   0.3453 |   0.2646 |   1.3048 |
-------------------------------------------------------------------------------

Target [ee1e70b] : BUG: DatetimeIndex cannot parse string ndarray with dayfirst
Base   [ecaa39c] : Merge pull request #7087 from jreback/series_cache

[jreback]

thanks!