Difference between loc() and iloc() in Pandas DataFrame

Let us understand the example given below:

# Importing pandas package
import pandas as pd

# Creating a dictionary of student marks
d = {
    "Peter":[65,70,70,75],
    "Harry":[45,56,66,66],
    "Tom":[67,87,65,53],
    "John":[56,78,65,64]
}

# Now, Create DataFrame and assign index 
# name as subject names
df = pd.DataFrame(d,index=["Maths","Physics","Chemistry","English"])

# Printing the DataFrame
print(df,"\n")

Output:

           Peter  Harry  Tom  John
Maths         65     45   67    56
Physics       70     56   87    78
Chemistry     70     66   65    65
English       75     66   53    64 

In the above example, we have created a DataFrame. Now we will use the loc() and iloc() methods and observer the difference between them.

Using loc() Method

The loc() method is a type of data selection method which takes the name of a row or column as a parameter. To perform various operations using the loc() method, we need to pass the required condition of rows and columns to get the filtered data.

# Selecting a row value where column name 
# is Harry and its value is 66
print(df.loc[df['Harry']==66],"\n")

# Selecting a row where column name 
# is Peter and its value is 70
print(df.loc[df['Peter']==70])

Output:

           Peter  Harry  Tom  John
Maths         65     45   67    56
Physics       70     56   87    78
Chemistry     70     66   65    65
English       75     66   53    64 

           Peter  Harry  Tom  John
Chemistry     70     66   65    65
English       75     66   53    64 

           Peter  Harry  Tom  John
Physics       70     56   87    78
Chemistry     70     66   65    65

Here, the loc() method returns the entire row where marks of 'Harry' and 'Peter' are 66 and 70 respectively. In this way, loc() method works upon a particular condition.

Using iloc() Method

The 'i' in iloc() stands for index. This is also a data selection method but here, we need to pass the proper index as a parameter to select the required row or column. Index are nothing but the integer value ranging from 0 to n-1 which represents the number of rows or columns. We can perform various operations using iloc() method. Inside iloc() method, the index value of the row comes first followed by the number of columns.

# Selecting 0th & 2nd index rows
print(df.iloc[[0, 2,]])

# Selecting specified rows and specified columns
print(df.iloc[0:3,1:2])

Output:

           Peter  Harry  Tom  John
Maths         65     45   67    56
Physics       70     56   87    78
Chemistry     70     66   65    65
English       75     66   53    64 

           Peter  Harry  Tom  John
Maths         65     45   67    56
Chemistry     70     66   65    65
           Harry
Maths         45
Physics       56
Chemistry     66

Here, first, we have passed a list of indices which means all those rows which we want to be displayed, and second time, we have passed the sliced index without any list, which means we want the rows from 0 to 2 followed by the columns number 1.

Note: The slicing method does not consider the end value i.e., the value specified after the colon (:).