Python Pandas：将行值重新用于另一行 - 跨行查找

Question

我有以下数据框

df1 =  DataFrame([['OBJ1', 10, 'BX', 'pool1', 'OBJ2'],['OBJ2', 0, '', '', 'OBJ1'],['OBJ3', 10, 'BY', 'pool2', 'OBJ4'],['OBJ4', 0, '', '', 'OBJ3'],['OBJ5', 10, 'BZ', 'pool3', '']], columns=['OBJ', 'value', 'conf', 'Res', 'Key'])

我想做的是：

当值为 0 时检查
Conf 和 Res 采用与 OBJ 匹配的键的行的值
例如OBJ2值为0，key为OBJ 2，则conf应为BX，Res为pool1

我使用查找或其他一些帖子尝试了多种解决方案，但似乎没有任何效果。

df1.loc[df1['value']==0, 'conf'] = df1.loc[df1['OBJ']==df1['Key']]['conf']

失败，因为我意识到这是在寻找具有 OBJ = Key

的行

Answer 1

使用DataFrame.merge with left_on and right_on parameters and index column created by DataFrame.reset_index for new DataFrame, then convert index column to index by DataFrame.set_index and last set new volumns by DataFrame.loc:

m = df1['value'].eq(0)
cols = ['conf','Res']
df = (df1.reset_index().loc[m, ['index','OBJ']]
          .merge(df1, left_on='OBJ', right_on='Key')
          .set_index('index')[cols])
print (df)
      conf    Res
index            
1       BX  pool1
3       BY  pool2

df1.loc[m, cols] = df
print (df1)
    OBJ  value conf    Res   Key
0  OBJ1     10   BX  pool1  OBJ2
1  OBJ2      0   BX  pool1  OBJ1
2  OBJ3     10   BY  pool2  OBJ4
3  OBJ4      0   BY  pool2  OBJ3
4  OBJ5     10   BZ  pool3

Python Pandas：将行值重新用于另一行 - 跨行查找

Python Pandas: reusing row value to another row - lookup across row

python

lookup

vlookup

pandas