示例#1
0
 def work(table, keys):
     df = data_utils.load_table(table,
                                usecols=keys,
                                dtype={k: pd.Int64Dtype() for k in keys})
     groupby_ss = df.groupby(keys).size()
     bct = groupby_ss.to_frame(name="weight").reset_index()
     for key in keys:
         kct = df.groupby(key).size().rename(f"{key}.cnt")
         bct = bct.merge(kct, how="left", left_on=key, right_index=True)
     return bct.astype(np.int64, copy=False)
示例#2
0
def load_data_table(table, join_keys):
    return data_utils.load_table(table,
                                 dtype={k: pd.Int64Dtype() for k in join_keys})
示例#3
0
def load_data_table(table, join_keys, usecols):
    return data_utils.load_table(table,
                                 usecols=usecols,
                                 dtype={k: np.int64
                                        for k in join_keys})