遍历所有数据帧列

Question

I want to compare all rows of 2 given dataframes

how can i optimize the following code to dynamically iterate through all columns of the given pandas dataframe?

df1,df2 = pd.read_csv(...)

for index2, row2 in df2.iterrows():
    for index1, row1 in df1.iterrows():
        if row1[0]==row2[0]: i = i+1
        if row1[1]==row2[1]: i = i+1
        if row1[2]==row2[2]: i = i+1
        if row1[3]==row2[3]: i = i+1
        print("# same values: "+str(i))
        i = 0

Answer 1

IIUC you need to check whether whole row of one dataframe is equal to another one. You could compare for equality two dataframes then use all method for that with axis=1 to check rows and then summing the result:

df1 = pd.DataFrame({'a': [1, 2, 3, 4, 5], 'b': [2, 3, 4, 5, 6]})
df2 = pd.DataFrame({'a': [1, 5, 3, 7, 5], 'b': [2, 3, 8, 5, 6]})

In [1531]: df1 == df2
Out[1531]: 
       a      b
0   True   True
1  False   True
2   True  False
3  False   True
4   True   True

In [1532]: (df1 == df2).all(axis=1)
Out[1532]: 
0     True
1    False
2    False
3    False
4     True
dtype: bool

In [1533]: (df1 == df2).all(axis=1).sum()
Out[1533]: 2

result = (df1 == df2).all(axis=1).sum()

In [1535]: print("# same values: "+str(result))
# same values: 2

Answer 2

Your nested for loop suggests that you are comparing all rows of the first DataFrame to all rows of the second DataFrame , and count the cases where values in corresponding columns match.

If so, you can fall back on numpy broadcasting to sum the equal cases for each row in df1 relative to all rows in df2 , and then sum these for all rows in df1 to get the total like so:

df1.apply(lambda x: np.sum(df2.values == x.values), axis=1)

To illustrate, two randomly sampled DataFrames :

df1 = pd.DataFrame(np.random.randint(1, 5, (10, 2)))

   0  1
0  2  4
1  2  3
2  4  1
3  3  3
4  3  3
5  4  4
6  2  4
7  3  4
8  3  4
9  4  1

df2 = pd.DataFrame(np.random.randint(1, 5, (10, 2)))

   0  1
0  3  2
1  3  4
2  4  4
3  2  3
4  4  3
5  4  1
6  4  1
7  3  4
8  3  1
9  1  4

Get the sum of equal values for all df1 rows after comparing each to all df2 rows :

df1.apply(lambda x: np.sum(df2.values == x.values), axis=1)

0    5
1    3
2    7
3    6
4    6
5    8
6    5
7    8
8    8
9    7

And you could then sum the cases, or do it all in one go:

df1.apply(lambda x: np.sum(df2.values == x.values), axis=1).sum()

63