Programmer's Picnic · Python / Pandas · Lesson 2

Change and Update DataFrames

Correct a value, modify only matching rows, calculate new columns and add or remove records without fragile chained assignment.

0 to infinity · Learn by making real DataFrames · No file required

Start from this table

Each exercise below rebuilds the same sample DataFrame, so changes in one example do not accidentally affect the next.

Row	Name	City	Marks	FeePaid
0	Asha	Varanasi	88	True
1	Ravi	Ranchi	42	False
2	Meera	Varanasi	76	True
3	Kabir	Patna	91	False

Python · pandas

import pandas as pd

students = pd.DataFrame({
    "Name": ["Asha", "Ravi", "Meera", "Kabir"],
    "City": ["Varanasi", "Ranchi", "Varanasi", "Patna"],
    "Marks": [88, 42, 76, 91],
    "FeePaid": [True, False, True, False]
})

Modern pandas rule: Edit a DataFrame using a single .loc[row_selection, column_selection] operation. Avoid expressions such as students["Marks"][mask] = 50 — chained assignment does not update the original under Copy-on-Write.

Step 1

Correct a specific student's marks

Select a row by condition and the column to edit in one step. Print the updated row to verify.

Python · pandas

import pandas as pd

students = pd.DataFrame({
    "Name": ["Asha", "Ravi", "Meera", "Kabir"],
    "City": ["Varanasi", "Ranchi", "Varanasi", "Patna"],
    "Marks": [88, 42, 76, 91],
    "FeePaid": [True, False, True, False]
})

students.loc[students["Name"] == "Ravi", "Marks"] = 52
print(students.loc[students["Name"] == "Ravi"])

Remember: This updates the original DataFrame directly and intentionally.

Step 2

Edit one cell with .at or .iat

Use .at when you already know the row and column label. Use .iat when you know their integer positions.

Python · pandas

import pandas as pd

students = pd.DataFrame({
    "Name": ["Asha", "Ravi", "Meera", "Kabir"],
    "City": ["Varanasi", "Ranchi", "Varanasi", "Patna"],
    "Marks": [88, 42, 76, 91],
    "FeePaid": [True, False, True, False]
})

students.at[0, "City"] = "Prayagraj"
students.iat[1, 2] = 49
print(students)

Remember: Use the right accessor for your indexing scheme; label 0 and position 0 only coincide with the default RangeIndex.

Step 3

Update many matching rows together

Use a mask to change several rows in a single operation, keeping the others unchanged.

Python · pandas

import pandas as pd

students = pd.DataFrame({
    "Name": ["Asha", "Ravi", "Meera", "Kabir"],
    "City": ["Varanasi", "Ranchi", "Varanasi", "Patna"],
    "Marks": [88, 42, 76, 91],
    "FeePaid": [True, False, True, False]
})

students.loc[students["Marks"] < 50, "Marks"] = 50
students.loc[students["City"] == "Varanasi", "FeePaid"] = True
print(students)

Remember: A conditional assignment is safer than manually searching and editing each matching row.

Step 4

Add calculated or descriptive columns

Assign a new column directly; its values are calculated row by row. Then update some statuses.

Python · pandas

import pandas as pd

students = pd.DataFrame({
    "Name": ["Asha", "Ravi", "Meera", "Kabir"],
    "City": ["Varanasi", "Ranchi", "Varanasi", "Patna"],
    "Marks": [88, 42, 76, 91],
    "FeePaid": [True, False, True, False]
})

students["Bonus"] = 5
students["FinalMarks"] = (students["Marks"] + students["Bonus"]).clip(upper=100)
students["Status"] = "On track"
students.loc[students["Marks"] < 50, "Status"] = "Needs practice"
print(students[["Name", "FinalMarks", "Status"]])

Remember: The new columns align with the existing rows.

Step 5

Replace text values and rename a column

You can produce a modified Series, then assign it back. Rename a column and retain the returned DataFrame.

Python · pandas

import pandas as pd

students = pd.DataFrame({
    "Name": ["Asha", "Ravi", "Meera", "Kabir"],
    "City": ["Varanasi", "Ranchi", "Varanasi", "Patna"],
    "Marks": [88, 42, 76, 91],
    "FeePaid": [True, False, True, False]
})

students["City"] = students["City"].replace({"Ranchi": "Ranchi City"})
students = students.rename(columns={"Marks": "Score"})
print(students)

Remember: Assign the result of operations such as rename() when you want to keep the change.

Step 6

Add another student

For a simple default row index, add a new row using a matching list of values. For large collections, gather records before constructing a DataFrame.

Python · pandas

import pandas as pd

students = pd.DataFrame({
    "Name": ["Asha", "Ravi", "Meera", "Kabir"],
    "City": ["Varanasi", "Ranchi", "Varanasi", "Patna"],
    "Marks": [88, 42, 76, 91],
    "FeePaid": [True, False, True, False]
})

students.loc[len(students)] = ["Nina", "Varanasi", 83, True]
print(students.tail(2))

Remember: Values must match the DataFrame's column order. This simple pattern assumes the default consecutive row index.

Step 7

Remove a row or column

Create a filtered DataFrame when removing rows, or use drop() for named columns. Assign the result to keep it.

Python · pandas

import pandas as pd

students = pd.DataFrame({
    "Name": ["Asha", "Ravi", "Meera", "Kabir"],
    "City": ["Varanasi", "Ranchi", "Varanasi", "Patna"],
    "Marks": [88, 42, 76, 91],
    "FeePaid": [True, False, True, False]
})

students = students.loc[students["Name"] != "Ravi"].copy()
students = students.drop(columns=["FeePaid"])
print(students)

Remember: A new filtered DataFrame has the original row labels until you explicitly reset_index(drop=True).

Step 8

Update with two conditions

Combine the search skills from Lesson 1 and the editing skills from Lesson 2.

Python · pandas

import pandas as pd

students = pd.DataFrame({
    "Name": ["Asha", "Ravi", "Meera", "Kabir"],
    "City": ["Varanasi", "Ranchi", "Varanasi", "Patna"],
    "Marks": [88, 42, 76, 91],
    "FeePaid": [True, False, True, False]
})

mask = (students["City"] == "Varanasi") & (students["Marks"] >= 75)
students.loc[mask, "Marks"] = (students.loc[mask, "Marks"] + 3).clip(upper=100)
print(students[["Name", "Marks"]])

Remember: The assignment targets the DataFrame once. This pattern works with pandas Copy-on-Write.

Practice C · Fix and reward marks

Correct Ravi's marks to 55, give all students scoring at least 80 a bonus of 5 (maximum 100), and display the final results.

Reveal a possible solution

Python · pandas

import pandas as pd

students = pd.DataFrame({
    "Name": ["Asha", "Ravi", "Meera", "Kabir"],
    "City": ["Varanasi", "Ranchi", "Varanasi", "Patna"],
    "Marks": [88, 42, 76, 91],
    "FeePaid": [True, False, True, False]
})

students.loc[students["Name"] == "Ravi", "Marks"] = 55
mask = students["Marks"] >= 80
students.loc[mask, "Marks"] = (students.loc[mask, "Marks"] + 5).clip(upper=100)
print(students[["Name", "Marks"]])

Capstone — one script from creation to correction

Here is the entire beginner workflow in one program: create → search → change → verify.

Python · pandas

import pandas as pd

students = pd.DataFrame({
    "Name": ["Asha", "Ravi", "Meera", "Kabir"],
    "City": ["Varanasi", "Ranchi", "Varanasi", "Patna"],
    "Marks": [88, 42, 76, 91],
    "FeePaid": [True, False, True, False]
})

# Find students who need help.
print("Before correction:")
print(students.loc[students["Marks"] < 50, ["Name", "Marks"]])

# Correct a score and give a small reward to high scorers.
students.loc[students["Name"] == "Ravi", "Marks"] = 52
students["Eligible"] = students["Marks"] >= 75

print("After correction:")
print(students[["Name", "Marks", "Eligible"]])

Official guidance: Copy-on-Write and chained assignment. Continue with Lesson 3: cleaning missing values, duplicates and invalid data.

← Previous lesson Course index Next lesson: DataFrame cleaning →