loc vs iloc mental model in Pandas

Introduction

We use loc and iloc to pick rows and columns from tables in pandas. They help us get exactly the data we want.

When you want to select data by row and column labels (names).

When you want to select data by row and column positions (numbers).

When you want to slice parts of a table using labels or positions.

When you want to filter data based on index or column names.

When you want to avoid mistakes by clearly choosing label-based or position-based selection.

Syntax

Pandas

df.loc[row_label, column_label]
df.iloc[row_position, column_position]

loc uses the actual names of rows and columns.

iloc uses the number positions starting from 0.

Examples

Selects the value where row label is 'row1' and column label is 'colA'.

Pandas

df.loc['row1', 'colA']

Selects the value at first row (position 0) and second column (position 1).

Pandas

df.iloc[0, 1]

Selects rows from 'row1' to 'row3' and columns from 'colA' to 'colC' inclusive.

Pandas

df.loc['row1':'row3', 'colA':'colC']

Selects first 3 rows and first 2 columns by position (end exclusive).

Pandas

df.iloc[0:3, 0:2]

Sample Program

This code shows how to pick single values and slices from a table using loc and iloc. It uses labels for loc and positions for iloc.

Pandas

import pandas as pd

# Create a simple DataFrame with labels
data = {'colA': [10, 20, 30], 'colB': [40, 50, 60], 'colC': [70, 80, 90]}
index_labels = ['row1', 'row2', 'row3']
df = pd.DataFrame(data, index=index_labels)

# Using loc to select by labels
val_loc = df.loc['row2', 'colB']

# Using iloc to select by positions
val_iloc = df.iloc[1, 1]

# Using loc to slice rows and columns by labels
slice_loc = df.loc['row1':'row2', 'colA':'colB']

# Using iloc to slice rows and columns by positions
slice_iloc = df.iloc[0:2, 0:2]

print('Value with loc:', val_loc)
print('Value with iloc:', val_iloc)
print('\nSlice with loc:')
print(slice_loc)
print('\nSlice with iloc:')
print(slice_iloc)

Important Notes

loc includes the last label in slices (end is inclusive).

iloc excludes the last position in slices (end is exclusive).

If your DataFrame index is numbers starting from 0, loc and iloc can give different results.

Summary

loc selects data by row and column labels.

iloc selects data by row and column positions (numbers starting at 0).

Use loc when you know the names, and iloc when you want to use positions.

Practice

1. What is the main difference between loc and iloc in pandas?

easy

A. loc uses labels, iloc uses integer positions

B. loc uses integer positions, iloc uses labels

C. loc selects columns only, iloc selects rows only

D. loc is faster than iloc

Solution

1. Step 1: Understand loc selection method

   loc selects data by row and column labels, meaning it uses the actual names of rows and columns.

2. Step 2: Understand iloc selection method

   iloc selects data by integer positions, starting from 0, ignoring labels.

3. Final Answer:

   loc uses labels, iloc uses integer positions -> Option A

4. Quick Check:

   loc = labels, iloc = positions [OK]

Hint: Remember: loc = labels, iloc = integer positions [OK]

Common Mistakes:

• Confusing labels with positions
• Thinking loc uses positions
• Assuming iloc uses labels

2. Which of the following is the correct syntax to select rows with labels 'a' to 'c' and columns 'x' to 'z' using loc?

easy

A. df.loc['a':'c', 'x':'z']

B. df.iloc['a':'c', 'x':'z']

C. df.loc[0:2, 0:2]

D. df.iloc[0:2, 'x':'z']

Solution

1. Step 1: Identify loc syntax

   loc uses labels, so slicing with labels like 'a':'c' and 'x':'z' is valid.

2. Step 2: Check other options

   Selections using iloc with label slices are invalid because iloc requires integer positions. Using integer slices with loc is incorrect when the labels are strings like 'a', 'c', etc.

3. Final Answer:

   df.loc['a':'c', 'x':'z'] -> Option A

4. Quick Check:

   loc uses labels in slices [OK]

Hint: Use labels inside loc brackets, positions inside iloc [OK]

Common Mistakes:

• Using integer positions with loc
• Using labels with iloc
• Mixing loc and iloc syntax

3. Given the DataFrame:

import pandas as pd
df = pd.DataFrame({'A':[10,20,30], 'B':[40,50,60]}, index=['x','y','z'])

What is the output of df.iloc[1, 0]?

medium

A. 30

B. 40

C. 50

D. 20

Solution

1. Step 1: Understand DataFrame structure

   Rows are labeled 'x','y','z' with values in column 'A' as [10,20,30]. Positions start at 0, so row 1 is 'y', column 0 is 'A'.

2. Step 2: Apply iloc selection

   df.iloc[1, 0] selects row at position 1 ('y') and column at position 0 ('A'), value is 20.

3. Final Answer:

   20 -> Option D

4. Quick Check:

   iloc uses positions: row 1, col 0 = 20 [OK]

Hint: iloc uses zero-based positions, count rows and columns from 0 [OK]

Common Mistakes:

• Using labels instead of positions
• Confusing row and column indices
• Selecting wrong cell by mixing loc and iloc

4. Consider this code:

import pandas as pd
df = pd.DataFrame({'A':[1,2,3]}, index=[10,20,30])
print(df.loc[1])

What error will occur and why?

medium

A. No error, prints 2

B. KeyError because label 1 does not exist in index

C. IndexError because position 1 is out of range

D. TypeError because loc expects string labels

Solution

1. Step 1: Check DataFrame index labels

   Index labels are [10, 20, 30], so label 1 is not present.

2. Step 2: Understand loc behavior

   loc looks for label 1, which does not exist, causing a KeyError.

3. Final Answer:

   KeyError because label 1 does not exist in index -> Option B

4. Quick Check:

   loc needs existing labels, else KeyError [OK]

Hint: loc needs exact labels; missing label causes KeyError [OK]

Common Mistakes:

• Thinking loc uses positions
• Expecting no error when label missing
• Confusing KeyError with IndexError

5. You have a DataFrame with non-numeric row labels:

import pandas as pd
df = pd.DataFrame({'Score':[88, 92, 95]}, index=['John', 'Jane', 'Doe'])

You want to select the second row using loc and iloc. Which code will correctly select the second row and what will be the difference in output?

hard

A. df.loc['Jane'] and df.iloc['Jane'] both select by label 'Jane'; output differs

B. df.loc[1] selects by position 1; df.iloc['Jane'] selects by label 'Jane'; both return the same row

C. df.loc['Jane'] selects by label 'Jane'; df.iloc[1] selects by position 1; both return the same row

D. df.loc[1] and df.iloc[1] both select by position 1; output differs

Solution

1. Step 1: Understand row labels and positions

   Rows are labeled 'John', 'Jane', 'Doe' at positions 0, 1, 2 respectively.

2. Step 2: Use loc and iloc correctly

   df.loc['Jane'] selects row with label 'Jane'. df.iloc[1] selects row at position 1, which is also 'Jane'. Both return the same row.

3. Final Answer:

   df.loc['Jane'] selects by label 'Jane'; df.iloc[1] selects by position 1; both return the same row -> Option C

4. Quick Check:

   loc uses label, iloc uses position; same row here [OK]

Hint: loc needs label, iloc needs position; same row if label matches position [OK]

Common Mistakes:

• Using integer with loc
• Using label with iloc
• Assuming loc and iloc always differ