比较 2 个 DataFrame 并删除不包含相应 ID 变量的行

Question

我需要比较 2 个 DataFrame 并删除不包含相应 ID 的行。例如考虑 df1 和 df2.

df1 = pd.DataFrame({'ID':[1,2,3,4],
                'Food':['Ham','Cheese','Egg','Bacon',],
                'Amount':[5,2,10,4,],
                })

df2 = pd.DataFrame({'ID':[1,2,4,5],
                'Food':['Ham','Cheese','Bacon','Chocolate Salty Balls'],
                'Amount':[6,7,15,5000],
                })

伪代码：

if df1['ID'] notin df2['ID']:
   df2['ID'].drop()

...反之亦然。该示例中的结果将是：

df1:

   ID    Food  Amount
0   1     Ham       5
1   2  Cheese       2
2   4   Bacon       4

df2:

   ID    Food  Amount
0   1     Ham       6
1   2  Cheese       7
2   4   Bacon      15

Answer 1

您可以使用布尔掩码：

msk1 = df1['ID'].isin(df2['ID'])
msk2 = df2['ID'].isin(df1['ID'])
df1 = df1[msk1]
df2 = df2[msk2]

或使用 set.intersection 和布尔索引：

common = set(df1['ID']).intersection(df2['ID'])
df1 = df1[df1['ID'].isin(common)]
df2 = df2[df2['ID'].isin(common)]

输出：

df1:

   ID    Food  Amount
0   1     Ham       5
1   2  Cheese       2
3   4   Bacon       4

df2:

   ID    Food  Amount
0   1     Ham       6
1   2  Cheese       7
2   4   Bacon      15

Answer 2

inner_join() 来自 datar 怎么样：


>>> from datar.all import f, inner_join, select
>>> import pandas as pd
>>> df1 = pd.DataFrame({'ID':[1,2,3,4],
...                 'Food':['Ham','Cheese','Egg','Bacon',],
...                 'Amount':[5,2,10,4,],
...                 })
>>> 
>>> df2 = pd.DataFrame({'ID':[1,2,4,5],
...                 'Food':['Ham','Cheese','Bacon','Chocolate Salty Balls'],
...                 'Amount':[6,7,15,5000],
...                 })
>>> inner_join(df1, df2 >> select(f.ID))
       ID     Food  Amount
  <int64> <object> <int64>
0       1      Ham       5
1       2   Cheese       2
2       4    Bacon       4
>>> inner_join(df2, df1 >> select(f.ID))
       ID     Food  Amount
  <int64> <object> <int64>
0       1      Ham       6
1       2   Cheese       7
2       4    Bacon      15

比较 2 个 DataFrame 并删除不包含相应 ID 变量的行

Compare 2 DataFrames and drop rows that do not contain corresponding ID variables

python

compare

dataframe

pandas

boolean-indexing