pandas (original) (raw)

can you run some asv's i am not sure this is really hit much, but maybe

       before           after         ratio
     [a975a754]       [229b5fc1]
     <master>         <fix-is-numeric-helper>
+      1.84±0.1ms       2.37±0.2ms     1.28  io.csv.ReadCSVCachedParseDates.time_read_csv_cached(False)
+     4.16±0.09ms       4.86±0.5ms     1.17  io.csv.ReadCSVFloatPrecision.time_read_csv_python_engine(',', '.', 'high')
-     6.88±0.07ms      6.18±0.09ms     0.90  io.hdf.HDFStoreDataFrame.time_store_info

SOME BENCHMARKS HAVE CHANGED SIGNIFICANTLY.
PERFORMANCE DECREASED.

-----

       before           after         ratio
     [a975a754]       [229b5fc1]
     <master>         <fix-is-numeric-helper>
-     6.88±0.03ms      6.07±0.04ms     0.88  io.hdf.HDFStoreDataFrame.time_store_info
-        24.4±1ms       16.2±0.2ms     0.66  io.hdf.HDFStoreDataFrame.time_read_store_table_wide

SOME BENCHMARKS HAVE CHANGED SIGNIFICANTLY.
PERFORMANCE INCREASED.

-----

BENCHMARKS NOT SIGNIFICANTLY CHANGED.

-----

       before           after         ratio
     [a975a754]       [229b5fc1]
     <master>         <fix-is-numeric-helper>
+      18.5±0.4ms       24.4±0.7ms     1.32  io.csv.ReadCSVThousands.time_thousands(',', None)
+      17.7±0.8ms       22.4±0.6ms     1.26  io.csv.ReadCSVThousands.time_thousands(',', ',')
+      17.9±0.6ms         21.8±1ms     1.22  io.csv.ReadCSVThousands.time_thousands('|', ',')
-     6.94±0.08ms      6.14±0.04ms     0.89  io.hdf.HDFStoreDataFrame.time_store_info
-     2.09±0.08ms      1.78±0.03ms     0.85  io.csv.ReadCSVFloatPrecision.time_read_csv(';', '_', 'high')
-        5.63±1ms      3.95±0.03ms     0.70  io.csv.ReadCSVFloatPrecision.time_read_csv_python_engine(',', '.', 'high')
-      23.6±0.7ms       16.1±0.6ms     0.68  io.hdf.HDFStoreDataFrame.time_read_store_table_wide

SOME BENCHMARKS HAVE CHANGED SIGNIFICANTLY.
PERFORMANCE DECREASED.

-----

BENCHMARKS NOT SIGNIFICANTLY CHANGED.

Ran the IO asvs a few times and results aren't consistent. Maybe there's something going on with io.csv / io.hdf but hard to say.