Pandas在列中找到相似的行，根据条件创建新列

Question

我有一个df， 查询主题相同
0 WP_77.1 WP_706.1 HPS_1
1 WP_78.1 WP_46.1 HPS_2
2 WP_57.1 WP_26.1 HPS_3
3 WP_57.1 WP_627.1 HPS_4
4 WP_15.1 WP_16.1 HPS_5
5 WP_15.1 WP_17.1 HPS_6
6 WP_15.1 WP_63.1 HPS_7
7 WP_15.1 WP_61.1 HPS_8
8 WP_15.1 WP_56.1 HPS_9
9水电站40.1水电站11.1水电站10

我试过了

df['query_s'] = df['query'].shift(-1)
df['HPSame_s'] = df['HPSame'].shift(-1)
condition = [(df['query'] == df['query_s'])]
ifTrue = df['HPSame']
ifFalse = df['HPSame_s']
df['match'] = np.where(condition, ifTrue, ifFalse)

这会引发ValueError：值的长度与索引的长度不匹配

我也试过了，但没有达到预期效果

df.loc[(df['query'] == df['query_s']), 'match'] = df['HPSame']
df.loc[(df['query'] != df['query_s']), 'match'] = df['HPSame_s']

我正在寻找的结果是， df= 查询主题相同匹配 0 WP_77.1 WP_706.1 HPS_1 HPS_1 1 WP_78.1 WP_46.1 HPS_2 HPS_2 2 WP_57.1 WP_26.1 HPS_3 HPS_3 3 WP_57.1 WP_627.1水电站4水电站3 4 WP_15.1 WP_16.1 HPS_5 HPS_5 5水电站15.1水电站17.1水电站6水电站5 6 WP_15.1 WP_63.1水电站7水电站5 7水电站15.1水电站61.1水电站8水电站5 8水电站15.1水电站56.1水电站9水电站5 9 WP_40.1 WP_11.1 HPS_10 HPS_10

Answer 1

您可以使用ffill：

df['match'] = df['HPSame'].where(df['query'] != df['query'].shift()).ffill()

输出：

     query   subject  HPSame   match
0  WP_77.1  WP_706.1   HPS_1   HPS_1
1  WP_78.1   WP_46.1   HPS_2   HPS_2
2  WP_57.1   WP_26.1   HPS_3   HPS_3
3  WP_57.1  WP_627.1   HPS_4   HPS_3
4  WP_15.1   WP_16.1   HPS_5   HPS_5
5  WP_15.1   WP_17.1   HPS_6   HPS_5
6  WP_15.1   WP_63.1   HPS_7   HPS_5
7  WP_15.1   WP_61.1   HPS_8   HPS_5
8  WP_15.1   WP_56.1   HPS_9   HPS_5
9  WP_40.1   WP_11.1  HPS_10  HPS_10

Answer 2

您还可以使用groupby.transform('first')，如中所示

df['match'] = (
    df.groupby('query')['HPSame'].transform('first')
    .reset_index(drop=True)
)

输出

     query   subject  HPSame   match
0  WP_77.1  WP_706.1   HPS_1   HPS_1
1  WP_78.1   WP_46.1   HPS_2   HPS_2
2  WP_57.1   WP_26.1   HPS_3   HPS_3
3  WP_57.1  WP_627.1   HPS_4   HPS_3
4  WP_15.1   WP_16.1   HPS_5   HPS_5
5  WP_15.1   WP_17.1   HPS_6   HPS_5
6  WP_15.1   WP_63.1   HPS_7   HPS_5
7  WP_15.1   WP_61.1   HPS_8   HPS_5
8  WP_15.1   WP_56.1   HPS_9   HPS_5
9  WP_40.1   WP_11.1  HPS_10  HPS_10