Handling duplicate column names in Data Analysis Python - Time & Space Complexity

Time Complexity: Handling duplicate column names

O(n)

Understanding Time Complexity

When working with data tables, sometimes columns have the same name. We want to understand how the time to fix these duplicates changes as the table grows.

How does the work needed to find and rename duplicates grow with more columns?

Scenario Under Consideration

Analyze the time complexity of the following code snippet.

import pandas as pd

def rename_duplicates(df):
    cols = df.columns.tolist()
    seen = {}
    for i, col in enumerate(cols):
        if col in seen:
            seen[col] += 1
            cols[i] = f"{col}_{seen[col]}"
        else:
            seen[col] = 0
    df.columns = cols
    return df
    

This code renames duplicate column names by adding a number suffix to make each name unique.

Identify Repeating Operations

Identify the loops, recursion, array traversals that repeat.

• Primary operation: A single loop over all column names.
• How many times: Once for each column in the list.

How Execution Grows With Input

As the number of columns grows, the loop runs once per column, doing a quick check and update each time.

Input Size (n)	Approx. Operations
10	About 10 checks and updates
100	About 100 checks and updates
1000	About 1000 checks and updates

Pattern observation: The work grows directly with the number of columns. Double the columns, double the work.

Final Time Complexity

Time Complexity: O(n)

This means the time to rename duplicates grows in a straight line with the number of columns.

Common Mistake

[X] Wrong: "Checking for duplicates inside the loop makes this code slower than linear time."

[OK] Correct: The dictionary lookup for seen columns is very fast and happens once per column, so the total work still grows linearly.

Interview Connect

Understanding how to handle duplicates efficiently shows you can work with real messy data and keep your code running smoothly as data grows.

Self-Check

"What if we used a nested loop to compare each column with every other column? How would the time complexity change?"

Practice

1. What happens if a pandas DataFrame has duplicate column names and you try to select one column by name using df['column_name']?

easy

A. It merges the columns with the same name into one.

B. It returns all columns with that name as a DataFrame.

C. It returns only the first column with that name.

D. It raises a KeyError.

Solution

1. Step 1: Understand pandas behavior with duplicate columns

   When selecting a column by name, pandas returns all columns with that name as a DataFrame if duplicates exist.

2. Step 2: Confirm selection result type

   Instead of a Series, pandas returns a DataFrame containing all matching columns.

3. Final Answer:

   It returns all columns with that name as a DataFrame. -> Option B

4. Quick Check:

   Duplicate columns selection = DataFrame [OK]

Hint: Selecting duplicate columns returns all as DataFrame, not error [OK]

Common Mistakes:

• Assuming it raises an error
• Thinking it returns only the first column
• Believing columns merge automatically

2. Which of the following is the correct way to rename duplicate columns in a pandas DataFrame df to make them unique?

easy

A. df.columns = pd.io.parsers.ParserBase({'names':df.columns})._maybe_dedup_names(df.columns)

B. df.columns = df.columns + '_1'

C. df.rename_duplicates()

D. df.columns = list(set(df.columns))

Solution

1. Step 1: Identify pandas method for deduplication

   pandas internal method _maybe_dedup_names can rename duplicate columns to unique names.

2. Step 2: Check each option

   A uses pandas internal deduplication correctly; B adds suffix to all columns, not just duplicates; C is invalid method; D removes duplicates losing data.

3. Final Answer:

   df.columns = pd.io.parsers.ParserBase({'names':df.columns})._maybe_dedup_names(df.columns) -> Option A

4. Quick Check:

   Use pandas internal deduplication = df.columns = pd.io.parsers.ParserBase({'names':df.columns})._maybe_dedup_names(df.columns) [OK]

Hint: Use pandas internal _maybe_dedup_names to rename duplicates [OK]

Common Mistakes:

• Trying to rename with set() which loses order and data
• Using non-existent method rename_duplicates()
• Adding suffix to all columns blindly

3. Given the DataFrame df with columns ['A', 'B', 'A'] and data:

   A  B  A
0  1  2  3
1  4  5  6

What is the output of df['A']?

medium

A. A Series with values [1, 4]

B. A DataFrame with columns 'A' and 'B'

C. A DataFrame with two columns both named 'A' containing values [[1, 3], [4, 6]]

D. Raises a KeyError

Solution

1. Step 1: Check columns and duplicates

   DataFrame has duplicate columns named 'A' at positions 0 and 2.

2. Step 2: Selecting df['A'] returns all columns named 'A'

   pandas returns a DataFrame with both 'A' columns, values [[1, 3], [4, 6]].

3. Final Answer:

   A DataFrame with two columns both named 'A' containing values [[1, 3], [4, 6]] -> Option C

4. Quick Check:

   Duplicate column selection = DataFrame with both columns [OK]

Hint: Selecting duplicate columns returns DataFrame with all matching columns [OK]

Common Mistakes:

• Expecting a single Series output
• Thinking it raises an error
• Confusing columns returned with all DataFrame columns

4. You have a DataFrame df with duplicate columns. You try to rename duplicates by running:

df.columns = df.columns + '_dup'

but it renames all columns, not just duplicates. How to fix this?

medium

A. Use df.columns = pd.io.parsers.ParserBase({'names':df.columns})._maybe_dedup_names(df.columns).

B. Drop duplicate columns before renaming.

C. Use df.rename(columns=lambda x: x+'_dup').

D. Use a loop to add suffix only to duplicate columns.

Solution

1. Step 1: Understand why all columns get renamed

   Adding '_dup' suffix to all columns renames every column, not just duplicates.

2. Step 2: Fix by renaming only duplicates

   Use a loop or comprehension to detect duplicates and rename only those columns.

3. Final Answer:

   Use a loop to add suffix only to duplicate columns. -> Option D

4. Quick Check:

   Rename duplicates selectively = loop method [OK]

Hint: Rename only duplicates by looping and checking counts [OK]

Common Mistakes:

• Renaming all columns blindly
• Using rename() without condition
• Dropping duplicates losing data

5. You have a DataFrame df with columns ['X', 'Y', 'X', 'Z', 'Y']. You want to rename duplicates by adding suffixes _1, _2, ... to make all column names unique. Which code correctly achieves this?

hard

A. df.columns = pd.Series(df.columns).groupby(df.columns).cumcount().add(1).astype(str).radd(df.columns + '_')

B. df.columns = pd.Series(df.columns).groupby(df.columns).cumcount().astype(str) + df.columns

C. df.columns = [f'{col}_{i}' for i, col in enumerate(df.columns)]

D. df.columns = list(set(df.columns))

Solution

1. Step 1: Understand the goal

   We want to add suffixes _1, _2, ... to duplicate columns to make names unique.

2. Step 2: Analyze each option

   A correctly groups by column names and adds suffixes starting at 1; B adds suffix before column name incorrectly; C adds suffixes by position, not duplicates; D removes duplicates losing data.

3. Final Answer:

   df.columns = pd.Series(df.columns).groupby(df.columns).cumcount().add(1).astype(str).radd(df.columns + '_') -> Option A

4. Quick Check:

   Groupby cumcount + suffix = unique names [OK]

Hint: Use groupby cumcount to add numeric suffixes to duplicates [OK]

Common Mistakes:

• Using set() which loses order and data
• Adding suffixes by position, not duplicates
• Ignoring duplicates and renaming all columns