How to Sort CSV by multiple columns in Python ?

In Python, to sort a CSV file by multiple columns, we can use the sort_values() method provided by the Python Pandas library. This method sorts DataFrame values by taking column names as arguments.

Common methods for sorting a CSV file by multiple columns include ?

• sort_values() with inplace ? Sort DataFrame by multiple columns, modifying the original

• sort_values() without inplace ? Sort DataFrame by multiple columns, creating a new DataFrame

• sort_index() ? Sort by index after sorting by specified columns

Using sort_values() with inplace

The sort_values() method with inplace=True directly modifies the original DataFrame ?

Example

Let's create a sample CSV data and sort it by multiple columns ?

import pandas as pd
import io

# Create sample CSV data
csv_data = """Car,Date_of_Purchase,Reg_Price
BMW,10/10/2020,1000
Lexus,10/12/2020,750
Audi,10/17/2020,750
Jaguar,10/16/2020,1500
Mustang,10/19/2020,1100
Lamborghini,10/22/2020,1000"""

# Read CSV data into DataFrame
dataFrame = pd.read_csv(io.StringIO(csv_data))

print("Original DataFrame:")
print(dataFrame)

# Sort by multiple columns: Reg_Price and Car
dataFrame.sort_values(["Reg_Price", "Car"], axis=0, ascending=True, inplace=True, na_position='first')

print("\nSorted DataFrame by Reg_Price and Car:")
print(dataFrame)

Original DataFrame:
          Car Date_of_Purchase  Reg_Price
0         BMW       10/10/2020       1000
1       Lexus       10/12/2020        750
2        Audi       10/17/2020        750
3      Jaguar       10/16/2020       1500
4     Mustang       10/19/2020       1100
5  Lamborghini       10/22/2020       1000

Sorted DataFrame by Reg_Price and Car:
          Car Date_of_Purchase  Reg_Price
2        Audi       10/17/2020        750
1       Lexus       10/12/2020        750
0         BMW       10/10/2020       1000
5  Lamborghini       10/22/2020       1000
4     Mustang       10/19/2020       1100
3      Jaguar       10/16/2020       1500

Using sort_values() without inplace

Using sort_values() without setting inplace=True returns a new sorted DataFrame, leaving the original unchanged ?

Example

import pandas as pd
import io

# Create sample CSV data
csv_data = """Car,Date_of_Purchase,Reg_Price
BMW,10/10/2020,1000
Lexus,10/12/2020,750
Audi,10/17/2020,750
Jaguar,10/16/2020,1500
Mustang,10/19/2020,1100
Lamborghini,10/22/2020,1000"""

# Read CSV data into DataFrame
dataFrame = pd.read_csv(io.StringIO(csv_data))

# Sort by multiple columns without modifying the original DataFrame
sortedDataFrame = dataFrame.sort_values(["Reg_Price", "Car"], axis=0, ascending=True, na_position='first')

print("Original DataFrame (unchanged):")
print(dataFrame)
print("\nSorted DataFrame:")
print(sortedDataFrame)

Original DataFrame (unchanged):
          Car Date_of_Purchase  Reg_Price
0         BMW       10/10/2020       1000
1       Lexus       10/12/2020        750
2        Audi       10/17/2020        750
3      Jaguar       10/16/2020       1500
4     Mustang       10/19/2020       1100
5  Lamborghini       10/22/2020       1000

Sorted DataFrame:
          Car Date_of_Purchase  Reg_Price
2        Audi       10/17/2020        750
1       Lexus       10/12/2020        750
0         BMW       10/10/2020       1000
5  Lamborghini       10/22/2020       1000
4     Mustang       10/19/2020       1100
3      Jaguar       10/16/2020       1500

Using sort_index()

The sort_index() method sorts a DataFrame by its index. After sorting by multiple columns, you can further sort by index to arrange rows in index order ?

Example

import pandas as pd
import io

# Create sample CSV data
csv_data = """Car,Date_of_Purchase,Reg_Price
BMW,10/10/2020,1000
Lexus,10/12/2020,750
Audi,10/17/2020,750
Jaguar,10/16/2020,1500
Mustang,10/19/2020,1100
Lamborghini,10/22/2020,1000"""

# Read CSV data into DataFrame
dataFrame = pd.read_csv(io.StringIO(csv_data))

# Sort by multiple columns first
sortedDataFrame = dataFrame.sort_values(["Reg_Price", "Car"], axis=0, ascending=True, na_position='first')

print("Sorted by columns:")
print(sortedDataFrame)

# Further sort by index to restore original index order
indexSortedDataFrame = sortedDataFrame.sort_index()

print("\nSorted by columns, then by index:")
print(indexSortedDataFrame)

Sorted by columns:
          Car Date_of_Purchase  Reg_Price
2        Audi       10/17/2020        750
1       Lexus       10/12/2020        750
0         BMW       10/10/2020       1000
5  Lamborghini       10/22/2020       1000
4     Mustang       10/19/2020       1100
3      Jaguar       10/16/2020       1500

Sorted by columns, then by index:
          Car Date_of_Purchase  Reg_Price
0         BMW       10/10/2020       1000
1       Lexus       10/12/2020        750
2        Audi       10/17/2020        750
3      Jaguar       10/16/2020       1500
4     Mustang       10/19/2020       1100
5  Lamborghini       10/22/2020       1000

Sort Parameters

Key parameters for sort_values() include ?

Conclusion

Use sort_values() with column names to sort CSV data by multiple columns. Set inplace=True to modify the original DataFrame, or omit it to create a new sorted DataFrame. Use sort_index() for additional index-based sorting.