Error with closed='left' and _adjust_bin_edges

[dalejung]

import pandas as pd

#8/6/12 is a Monday
ind = pd.DatetimeIndex(start="8/6/2012", end="8/26/2012", freq="D")
n = len(ind)
data = [[x] * 5 for x in range(n)]
df = pd.DataFrame(data, columns=['open', 'high', 'low', 'close', 'vol'], index=ind)

df.resample('W-MON', how='first', closed='left', label='left')

Will result in an error. I tracked it down to

      def _adjust_bin_edges(self, binner, ax_values):
        # Some hacks for > daily data, see #1471, #1458, #1483

        bin_edges = binner.asi8

        if self.freq != 'D' and is_superperiod(self.freq, 'D'):
            day_nanos = _delta_to_nanoseconds(timedelta(1))
            if self.closed == 'right':
                bin_edges = bin_edges + day_nanos - 1
            else:
                bin_edges = bin_edges + day_nanos

I'm not sure why the day is added when closed='left'. Even on datasets that don't start on a Monday, and thus don't error, it'll cause the intervals to start on Tuesday.

[wesm]

Thanks, dale. You were indeed correct that the bin adjustment logic was incorrect.