When to use Hadoop in modern data stacks - Time & Space Complexity

Time Complexity: When to use Hadoop in modern data stacks

O(n)

Understanding Time Complexity

We want to understand how the time it takes to run Hadoop jobs grows as data size increases.

This helps us know when Hadoop is a good choice in modern data setups.

Scenario Under Consideration

Analyze the time complexity of a simple Hadoop MapReduce job that counts words.

// Mapper function
map(key, value) {
  for each word in value {
    emit(word, 1);
  }
}

// Reducer function
reduce(word, counts) {
  sum = 0;
  for each count in counts {
    sum += count;
  }
  emit(word, sum);
}

This job reads text data, counts how many times each word appears, and outputs totals.

Identify Repeating Operations

Look at the loops that repeat work:

• Primary operation: The mapper loops over every word in the input data.
• How many times: Once for each word in the entire dataset.
• Secondary operation: The reducer loops over counts for each unique word.
• How many times: Once for each occurrence of that word.

How Execution Grows With Input

As the input data grows, the number of words grows roughly in proportion.

Input Size (n words)	Approx. Operations
10	About 10 map operations + reduce sums
100	About 100 map operations + reduce sums
1000	About 1000 map operations + reduce sums

Pattern observation: The work grows roughly in direct proportion to the number of words.

Final Time Complexity

Time Complexity: O(n)

This means the time to run the job grows linearly with the size of the input data.

Common Mistake

[X] Wrong: "Hadoop always runs fast regardless of data size because it is distributed."

[OK] Correct: Hadoop splits work across machines, but total work still grows with data size, so bigger data means longer total processing time.

Interview Connect

Understanding how Hadoop scales helps you explain when to choose it for big data tasks versus faster tools for smaller data.

Self-Check

"What if we changed the job to run only on a sample of the data? How would the time complexity change?"

Practice

1. Which scenario is best suited for using Hadoop in a modern data stack?

easy

A. Running quick, real-time analytics on small data

B. Performing basic calculations in a spreadsheet

C. Storing simple, small CSV files on a single machine

D. Processing very large datasets that don't fit on one computer

Solution

1. Step 1: Understand Hadoop's strength

   Hadoop is designed to handle very large datasets that require distributed storage and parallel processing.

2. Step 2: Match scenario to Hadoop's use case

   Processing data too big for one computer fits Hadoop's design, unlike small or real-time data tasks.

3. Final Answer:

   Processing very large datasets that don't fit on one computer -> Option D

4. Quick Check:

   Big data needs Hadoop [OK]

Hint: Choose Hadoop for huge data, not small or fast tasks [OK]

Common Mistakes:

• Thinking Hadoop is good for real-time analytics
• Using Hadoop for small datasets
• Confusing Hadoop with simple file storage

2. Which of the following is the correct way to describe Hadoop's role in data processing?

easy

A. Hadoop stores and processes data in a distributed way across many machines

B. Hadoop processes data only on a single machine

C. Hadoop is mainly used for real-time data streaming

D. Hadoop is a tool for creating small databases on laptops

Solution

1. Step 1: Recall Hadoop architecture

   Hadoop uses distributed storage (HDFS) and parallel processing (MapReduce) across many machines.

2. Step 2: Eliminate incorrect options

   Options B, C, and D describe Hadoop incorrectly; it is not single-machine, real-time streaming, or for small databases.

3. Final Answer:

   Hadoop stores and processes data in a distributed way across many machines -> Option A

4. Quick Check:

   Distributed storage and processing = Hadoop [OK]

Hint: Hadoop = distributed storage + processing [OK]

Common Mistakes:

• Thinking Hadoop runs on one machine only
• Confusing Hadoop with streaming tools like Kafka
• Assuming Hadoop is for small-scale data

3. Consider this code snippet that checks data size before choosing processing method:

data_size_gb = 500
if data_size_gb > 100:
    method = 'Hadoop'
else:
    method = 'Pandas'
print(method)

What will be the output?

medium

A. Error

B. Pandas

C. Hadoop

D. null

Solution

1. Step 1: Evaluate the condition

   data_size_gb is 500, which is greater than 100, so the if condition is true.

2. Step 2: Determine assigned method

   Since condition is true, method is set to 'Hadoop'.

3. Final Answer:

   Hadoop -> Option C

4. Quick Check:

   500 > 100 -> Hadoop [OK]

Hint: Check if data size > 100GB to pick Hadoop [OK]

Common Mistakes:

• Choosing Pandas for large data
• Confusing > with < in condition
• Expecting syntax error due to print

4. This code tries to decide when to use Hadoop but has a bug:

data_size = 50
if data_size >= 100:
    tool = 'Hadoop'
else:
    tool = 'Spark'
print(tool)

What is the bug if the goal is to use Hadoop only for data larger than 100GB?

medium

A. The condition uses >= instead of >, so Hadoop runs at 100GB

B. The else block should assign 'Hadoop' instead of 'Spark'

C. The print statement is missing parentheses

D. Variable data_size should be a string, not integer

Solution

1. Step 1: Analyze the condition

   The condition uses >= 100, so Hadoop runs when data_size is exactly 100, but goal is strictly greater than 100.

2. Step 2: Identify correct condition

   Changing >= to > fixes the bug to run Hadoop only for data sizes greater than 100.

3. Final Answer:

   The condition uses >= instead of >, so Hadoop runs at 100GB -> Option A

4. Quick Check:

   Use > for strictly greater than 100 [OK]

Hint: Use >, not >=, for strictly greater than checks [OK]

Common Mistakes:

• Confusing >= with > in conditions
• Thinking print() is wrong syntax
• Assuming data_size must be string

5. You have a data pipeline that processes 10TB of logs daily. The logs are unstructured and need batch processing overnight. Which modern data stack choice fits best?

hard

A. Use a real-time streaming tool only

B. Use Hadoop for distributed storage and batch processing

C. Use a single SQL database on one server

D. Use Excel to analyze the logs

Solution

1. Step 1: Understand data size and type

   10TB is very large and logs are unstructured, needing batch processing overnight.

2. Step 2: Match tool to requirements

   Hadoop handles large, unstructured data with batch processing well; others are unsuitable for scale or type.

3. Final Answer:

   Use Hadoop for distributed storage and batch processing -> Option B

4. Quick Check:

   Big unstructured batch data -> Hadoop [OK]

Hint: Big, unstructured, batch data? Pick Hadoop [OK]

Common Mistakes:

• Choosing single server for huge data
• Using real-time tools for batch needs
• Trying Excel for terabytes of data