将逻辑应用于熊猫的groupby的多行

Consider a data frame df

    date      time   isopen    isclose  openlate   closeearly
    20200201  0920   Y         N        Y          N
    20200201  1645   N         Y        N          Y 
    20200202  0900   Y         N        N          N
    20200202  1650   N         Y        N          Y 
    20200203  0910   Y         N        Y          N
    20200203  1700   N         Y        N          N 

openlate and closeearly were computed from time and isopen/isclose columns.

我想按日期展平，有类似

    date      openlate   closeearly
    20200201  Y          Y
    20200202  N          Y 
    20200203  Y          N 

Basically to collapse multiple rows by a groupby, partition or window function, then choose a logic which value to keep among the rows. I do have this situation on many occasions.

    df.groupby("date")..['openlate' take 'Y' if any df["openlate"] in group is 'Y', else 'N';
 'closeearly' takes 'Y' if any df["closeearly"] in the group is 'Y', else 'N']