Adding & Modifying

DataFrames are size-mutable— you can grow them. There's one rule that runs through all of it: assign to a name that exists and you modify it; assign to a new name and you create it.

1Adding & changing a column

Assign to df['col']. If col doesn't exist, a new column appears; if it does, its values change. A single value fills the whole column; a list sets each row.

df_columns.py

import pandas as pd

df = pd.DataFrame({
    'Name':    ['Rinku', 'Ritu', 'Pankaj'],
    'English': [65, 77, 72],
    'IP':      [89, 83, 95],
}, index=['ID 1', 'ID 2', 'ID 3'])

# New column, single value -> every row gets 79
df['Hindi'] = 79

# New column, a list -> one value per row
df['French'] = [98, 77, 63]

# 'IP' already exists -> this MODIFIES it
df['IP'] = df['IP'] + 5   # bonus marks!

print(df)

import pandas as pd

df = pd.DataFrame({
    'Name':    ['Rinku', 'Ritu', 'Pankaj'],
    'English': [65, 77, 72],
    'IP':      [89, 83, 95],
}, index=['ID 1', 'ID 2', 'ID 3'])

# New column, single value -> every row gets 79
df['Hindi'] = 79

# New column, a list -> one value per row
df['French'] = [98, 77, 63]

# 'IP' already exists -> this MODIFIES it
df['IP'] = df['IP'] + 5   # bonus marks!

print(df)

Watch Out

If you give a list, its length must match the number of rows — otherwise Pandas raises a ValueError.

2Adding & changing a row

Use .loc (or .at) with a row label and : for all columns. A missing label creates a new row; an existing one is overwritten.

df_rows.py

import pandas as pd

df = pd.DataFrame({
    'Name':    ['Rinku', 'Ritu', 'Pankaj'],
    'English': [65, 77, 72],
    'IP':      [89, 83, 95],
}, index=['ID 1', 'ID 2', 'ID 3'])

# 'ID 4' doesn't exist -> a NEW row is created
df.loc['ID 4'] = ['Ajay', 81, 99]

# 'ID 2' exists -> this row is MODIFIED
df.loc['ID 2'] = ['Ritu K.', 80, 85]

print(df)

import pandas as pd

df = pd.DataFrame({
    'Name':    ['Rinku', 'Ritu', 'Pankaj'],
    'English': [65, 77, 72],
    'IP':      [89, 83, 95],
}, index=['ID 1', 'ID 2', 'ID 3'])

# 'ID 4' doesn't exist -> a NEW row is created
df.loc['ID 4'] = ['Ajay', 81, 99]

# 'ID 2' exists -> this row is MODIFIED
df.loc['ID 2'] = ['Ritu K.', 80, 85]

print(df)

3Modifying a single cell

To change just one value, target the exact cell. .at (labels) and .iat (positions) are the clearest, but the df['col']['row'] shortcut works too.

df_cell.py

import pandas as pd

df = pd.DataFrame({
    'Name':    ['Rinku', 'Ritu', 'Pankaj'],
    'English': [65, 77, 72],
    'IP':      [89, 83, 95],
}, index=['ID 1', 'ID 2', 'ID 3'])

df.at['ID 3', 'English'] = 88     # by labels
df.iat[0, 2] = 91                 # by positions (row 0, col 2 = IP)

print(df)

import pandas as pd

df = pd.DataFrame({
    'Name':    ['Rinku', 'Ritu', 'Pankaj'],
    'English': [65, 77, 72],
    'IP':      [89, 83, 95],
}, index=['ID 1', 'ID 2', 'ID 3'])

df.at['ID 3', 'English'] = 88     # by labels
df.iat[0, 2] = 91                 # by positions (row 0, col 2 = IP)

print(df)

Key Takeaway

The golden rule: assigning to an existing column/row modifies it; assigning to a new label creates it. A single value broadcasts to all cells; a list sets them one by one.

Quick Check

df has no 'Science' column. What does df['Science'] = 50 do?

Quick Check

Which changes just ONE cell — Pankaj's IP mark?

← Previous

🎯 Rows & Subsets: loc / iloc

Deleting & Renaming ✏️