iloc for position-based selection in Pandas - Time & Space Complexity

Time Complexity: iloc for position-based selection

O(r x c)

Understanding Time Complexity

We want to understand how the time it takes to select data using iloc changes as the data size grows.

How does the number of rows or columns affect the work done when using iloc?

Scenario Under Consideration

Analyze the time complexity of the following code snippet.

import pandas as pd

df = pd.DataFrame({
    'A': range(1000),
    'B': range(1000, 2000),
    'C': range(2000, 3000)
})

subset = df.iloc[100:200, 1:3]

This code creates a DataFrame with 1000 rows and 3 columns, then selects rows 100 to 199 and columns 1 to 2 using iloc.

Identify Repeating Operations

Identify the loops, recursion, array traversals that repeat.

• Primary operation: Accessing each cell in the selected slice of the DataFrame.
• How many times: Once for each row and column in the selected range.

How Execution Grows With Input

When you select more rows or columns, the work grows with the size of the slice.

Input Size (rows x columns)	Approx. Operations
10 x 2	20
100 x 2	200
100 x 3	300

Pattern observation: The operations grow roughly in direct proportion to the number of cells selected.

Final Time Complexity

Time Complexity: O(r x c)

This means the time grows proportionally to the number of rows (r) times the number of columns (c) you select.

Common Mistake

[X] Wrong: "Selecting data with iloc always takes the same time no matter how much data is selected."

[OK] Correct: The time depends on how many rows and columns you pick because iloc accesses each cell in the selection.

Interview Connect

Understanding how data selection scales helps you write efficient code and explain your choices clearly in real projects and interviews.

Self-Check

What if we changed the selection to only one column but many rows? How would the time complexity change?

Practice

1. What does df.iloc[2] select from a pandas DataFrame df?

easy

A. The third row of the DataFrame by position

B. The row with index label 2

C. The third column of the DataFrame by position

D. The column with index label 2

Solution

1. Step 1: Understand iloc usage

   iloc selects rows or columns by integer position, starting at 0.

2. Step 2: Interpret df.iloc[2]

   This selects the row at position 2, which is the third row counting from zero.

3. Final Answer:

   The third row of the DataFrame by position -> Option A

4. Quick Check:

   iloc uses position, so 2 means third row [OK]

Hint: Remember: iloc counts rows/columns from zero position [OK]

Common Mistakes:

• Confusing iloc with loc (label-based selection)
• Thinking iloc uses index labels instead of positions
• Mixing row and column selection

2. Which of the following is the correct syntax to select the first 3 rows and first 2 columns using iloc?

easy

A. df.iloc[3, 2]

B. df.iloc[1:3, 1:2]

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

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

Solution

1. Step 1: Understand slicing with iloc

   Slices in iloc are start:stop and select positions from start up to but not including stop.

2. Step 2: Apply slice for first 3 rows and 2 columns

   First 3 rows are positions 0,1,2 so slice is :3. First 2 columns are positions 0,1 so slice is :2.

3. Final Answer:

   df.iloc[:3, :2] -> Option D

4. Quick Check:

   Slices exclude stop, so :3 means 0,1,2 [OK]

Hint: Use slice :n to get first n positions [OK]

Common Mistakes:

• Using wrong slice ranges that exclude needed rows/columns
• Confusing row and column order in iloc
• Using single integers instead of slices for multiple rows/columns

3. Given the DataFrame:

import pandas as pd
df = pd.DataFrame({
  'A': [10, 20, 30],
  'B': [40, 50, 60],
  'C': [70, 80, 90]
})
print(df.iloc[1, 2])

What is the output?

medium

A. 50

B. 30

C. 80

D. 60

Solution

1. Step 1: Identify row and column positions

   iloc[1, 2] means row at position 1 (second row), column at position 2 (third column).

2. Step 2: Find value at row 1, column 2

   Row 1 values: A=20, B=50, C=80. Column 2 is 'C', so value is 80.

3. Final Answer:

   80 -> Option C

4. Quick Check:

   Row 1, Col 2 = 80 [OK]

Hint: Remember iloc[row, column] uses zero-based positions [OK]

Common Mistakes:

• Mixing up row and column positions
• Using index labels instead of positions
• Selecting wrong column number

4. What is wrong with this code snippet?

df.iloc[1:3, 2:1]

It aims to select rows 1 and 2 and columns 2 and 1.

medium

A. iloc cannot slice columns, only rows

B. Column slice 2:1 is empty because stop is less than start

C. Row slice 1:3 is invalid syntax

D. The code will select columns 1 and 2 correctly

Solution

1. Step 1: Analyze column slice 2:1

   Slices select positions from start up to but not including stop. Here start=2, stop=1, which is reversed and results in an empty selection.

2. Step 2: Check row slice 1:3

   Row slice 1:3 selects rows at positions 1 and 2 correctly.

3. Final Answer:

   Column slice 2:1 is empty because stop is less than start -> Option B

4. Quick Check:

   Slices with stop less than start return empty [OK]

Hint: Slice stop must be greater than start to select data [OK]

Common Mistakes:

• Assuming reversed slices select columns in reverse
• Thinking iloc syntax is invalid for slices
• Confusing iloc with loc label slicing

5. You have a DataFrame df with 5 rows and 4 columns. You want to select the last 2 rows and the last 3 columns using iloc. Which code correctly does this?

hard

A. df.iloc[2:, 1:]

B. df.iloc[2:4, 1:3]

C. df.iloc[-2:, -2:]

D. df.iloc[3:5, 1:4]

Solution

1. Step 1: Understand row positions for last 2 rows

   With 5 rows (positions 0 to 4), last 2 rows are positions 3 and 4, so slice is 3:5 or equivalently -2:.

2. Step 2: Understand column positions for last 3 columns

   With 4 columns (positions 0 to 3), last 3 columns are positions 1, 2, 3, so slice is 1:4 or equivalently -3:.

3. Step 3: Match correct iloc syntax

   Using negative indices, df.iloc[-2:, -3:] would be correct. Among options, df.iloc[2:, 1:] selects rows from position 2 to end (rows 2,3,4) and columns from position 1 to end (columns 1,2,3), which includes the last 2 rows and last 3 columns.

4. Final Answer:

   df.iloc[2:, 1:] -> Option A

5. Quick Check:

   Rows 2-4 and columns 1-3 selected correctly [OK]

Hint: Count positions from zero; slice stop is one past last index [OK]

Common Mistakes:

• Using negative slices incorrectly
• Off-by-one errors in slice ranges
• Confusing row and column slice order