JIYIK CN >

Python Pandas DataFrame.transform() DataFrameapplies a function on and transforms DataFrame. The function to be applied is passed as an argument to the function. The axis length of transform()the transformed should be the same as the original .DataFrameDataFrame

pandas.DataFrame.transform()Syntax

DataFrame.transform(func, axis, *args, **kwargs)

parameter

Return Value

It returns a converted DataFramewhich has the same length as the original " DataFrame". If the returned DataFramehas unequal lengths then the function raises ValueError.

Sample code:DataFrame.transform()

Let's try this function first, DataFrameadding a number to each value of .

import pandas as pd

dataframe = pd.DataFrame({
                            'A': 
                                {0: 6, 
                                1: 20, 
                                2: 80,
                                3: 78,
                                4: 95}, 
                            'B': 
                                {0: 60, 
                                1: 50, 
                                2: 7,
                                3: 67,
                                4: 54}
                        })

print(dataframe)

Examples DataFrameare:

    A   B
0   6  60
1  20  50
2  80   7
3  78  67
4  95  54
5  98  34

This function has only one required parameter, func. Now, we will use this function DataFrameto add 20 to each value of .

import pandas as pd

dataframe = pd.DataFrame(
    {"A": {0: 6, 1: 20, 2: 80, 3: 78, 4: 95}, "B": {0: 60, 1: 50, 2: 7, 3: 67, 4: 54}}
)

dataframe1 = dataframe.transform(func=lambda x: x + 20)
print(dataframe1)

Output:

     A   B
0   26  80
1   40  70
2  100  27
3   98  87
4  115  74
5  118  54

lambdaThe keyword is used to declare an anonymous addition function.

Example code: DataFrame.transform()Using sqrta string as a function

import pandas as pd

dataframe = pd.DataFrame(
    {"A": {0: 6, 1: 20, 2: 80, 3: 78, 4: 95}, "B": {0: 60, 1: 50, 2: 7, 3: 67, 4: 54}}
)

dataframe1 = dataframe.transform(func="sqrt")
print(dataframe1)

Output:

          A         B
0  2.449490  7.745967
1  4.472136  7.071068
2  8.944272  2.645751
3  8.831761  8.185353
4  9.746794  7.348469
5  9.899495  5.830952

Here, lambdainstead of passing a function, we pass the function name as a string.

Example Code: DataFrame.transform()Passing a List of Functions

import pandas as pd

dataframe = pd.DataFrame(
    {"A": {0: 6, 1: 20, 2: 80, 3: 78, 4: 95}, "B": {0: 60, 1: 50, 2: 7, 3: 67, 4: 54}}
)

dataframe1 = dataframe.transform(func=["sqrt", "exp"])
print(dataframe1)

Output:

          A                       B              
       sqrt           exp      sqrt           exp
0  2.449490  4.034288e+02  7.745967  1.142007e+26
1  4.472136  4.851652e+08  7.071068  5.184706e+21
2  8.944272  5.540622e+34  2.645751  1.096633e+03
3  8.831761  7.498417e+33  8.185353  1.252363e+29
4  9.746794  1.811239e+41  7.348469  2.830753e+23

We passed a "list" of two function names ['sqrt', 'exp']as func. The returned DataFramecontains two extra columns because there is an extra function.

DataFrame.apply()vs DataFrame.transform()Function

We can also use DataFrame.apply() function to achieve the above results. But if we compare the two functions, we can see that DataFrame.transform()the function is more efficient when handling complex operations.