has2k1 / plotnine / 584 / 1

87%

master: 87%

push

travis-ci

Fix stat_count when dropping some limits

Use pd.isna instead of np.isnan when calculating the resolution

   np.isnan(np.array([1, 1], dtype=object))  # Now fails
   pd.isna(np.array([1, 1], dtype=object))   # passes

Reason: When we introduce NaNs for the dropped values, we change that
column to dtype=object (numpy).

2465 of 3166 branches covered (77.86%)

Branch coverage included in aggregate %.

8380 of 9290 relevant lines covered (90.2%)

0.9 hits per line