Join algorithms (nested loop, sort-merge, hash join) in DBMS Theory - Time & Space Complexity

Time Complexity: Join algorithms (nested loop, sort-merge, hash join)

O(n^2) for Nested Loop, O(n log n) for Sort-Merge, O(n) for Hash Join

Understanding Time Complexity

When databases combine tables using joins, the method chosen affects how long it takes. Understanding time complexity helps us see how the work grows as tables get bigger.

We want to know: How does the time to join tables change as the number of rows increases?

Scenario Under Consideration

Analyze the time complexity of these three common join methods.

-- Nested Loop Join
FOR each row in TableA LOOP
  FOR each row in TableB LOOP
    IF join condition matches THEN
      output joined row;
    END IF;
  END LOOP;
END LOOP;

-- Sort-Merge Join
Sort TableA and TableB on join key;
Merge rows by scanning both sorted tables once;

-- Hash Join
Build hash table on smaller table using join key;
Probe hash table with rows from larger table;

These snippets show how each join method processes rows to combine tables.

Identify Repeating Operations

Look at the main repeated steps in each join:

• Nested Loop Join: Two loops, one inside the other, checking every pair of rows.
• Sort-Merge Join: Sorting both tables, then one pass through both sorted lists.
• Hash Join: Building a hash table from one table, then checking each row of the other table against it.

How Execution Grows With Input

Imagine both tables have n rows:

Input Size (n)	Approx. Operations
10	Nested Loop: 100, Sort-Merge: ~66 + 20, Hash Join: ~10 + 10
100	Nested Loop: 10,000, Sort-Merge: ~1,300 + 200, Hash Join: ~100 + 100
1000	Nested Loop: 1,000,000, Sort-Merge: ~20,000 + 2,000, Hash Join: ~1000 + 1000

Nested loops grow very fast as tables get bigger, while sort-merge and hash join grow more slowly, roughly doubling or slightly more.

Final Time Complexity

Time Complexity: O(n^2) for Nested Loop Join, O(n log n) for Sort-Merge Join, and O(n) for Hash Join

This means nested loops take much longer as tables grow, sorting takes more time but less than nested loops, and hashing is usually fastest for large tables.

Common Mistake

[X] Wrong: "All join methods take the same time regardless of table size."

[OK] Correct: Different join methods handle data differently, so their time grows at different rates as tables get bigger.

Interview Connect

Knowing how join methods scale helps you explain database performance clearly. This skill shows you understand how data size affects query speed, a key part of working with databases.

Self-Check

What if one table is much smaller than the other? How would that affect the time complexity of each join method?

Practice

1. Which join algorithm compares each row of one table with every row of another table to find matching pairs?

easy

A. Index join

B. Sort-merge join

C. Hash join

D. Nested loop join

Solution

1. Step 1: Understand the nested loop join process

   Nested loop join works by taking each row from the first table and comparing it with every row in the second table to find matches.

2. Step 2: Compare with other join types

   Sort-merge join sorts and merges, hash join uses hashing, and index join uses indexes, so they do not compare every pair.

3. Final Answer:

   Nested loop join -> Option D

4. Quick Check:

   Every row compared = Nested loop join [OK]

Hint: Nested loop = check all pairs one by one [OK]

Common Mistakes:

• Confusing nested loop with hash join
• Thinking sort-merge compares all pairs
• Assuming index join is same as nested loop

2. Which of the following is the correct description of a hash join algorithm?

easy

A. Sort both tables and merge matching rows

B. Use a hash table to find matching rows quickly

C. Compare each row of one table with every row of another

D. Use indexes to speed up join operations

Solution

1. Step 1: Recall hash join working

   Hash join builds a hash table on one table's join key to quickly find matching rows from the other table.

2. Step 2: Eliminate other options

   Sorting and merging is sort-merge join, comparing all pairs is nested loop, and using indexes is index join, not hash join.

3. Final Answer:

   Use a hash table to find matching rows quickly -> Option B

4. Quick Check:

   Hash table = fast matching [OK]

Hint: Hash join uses hash table for fast lookup [OK]

Common Mistakes:

• Mixing hash join with sort-merge join
• Thinking hash join compares all pairs
• Confusing index usage with hash join

3. Consider two tables A and B, each with 1000 rows. Which join algorithm is likely to perform best if both tables are already sorted on the join key?

medium

A. Nested loop join

B. Hash join

C. Sort-merge join

D. Cartesian join

Solution

1. Step 1: Analyze the condition of sorted tables

   Since both tables are sorted on the join key, sort-merge join can efficiently merge them by scanning once through both tables.

2. Step 2: Compare with other algorithms

   Nested loop join checks all pairs (slow), hash join builds hash table (extra work), Cartesian join is unrelated and very expensive.

3. Final Answer:

   Sort-merge join -> Option C

4. Quick Check:

   Sorted tables = sort-merge join best [OK]

Hint: Sorted tables? Use sort-merge join [OK]

Common Mistakes:

• Choosing nested loop for sorted tables
• Assuming hash join is always fastest
• Confusing Cartesian join with normal join

4. A database query uses a hash join but runs very slowly. Which of the following is a likely cause?

medium

A. The hash table does not fit in memory causing disk spills

B. The join keys are unique in both tables

C. Both tables are already sorted

D. The nested loop join was used instead

Solution

1. Step 1: Understand hash join memory use

   Hash join builds a hash table in memory. If it is too large, it spills to disk, slowing performance.

2. Step 2: Evaluate other options

   Sorted tables do not slow hash join, unique keys help performance, and nested loop join is a different algorithm.

3. Final Answer:

   The hash table does not fit in memory causing disk spills -> Option A

4. Quick Check:

   Memory overflow = slow hash join [OK]

Hint: Hash join slow? Check memory size for hash table [OK]

Common Mistakes:

• Blaming sorted tables for hash join slowness
• Ignoring memory limits in hash join
• Confusing nested loop join with hash join

5. You have two large tables to join on a key. Table A fits in memory but Table B is very large and unsorted. Which join algorithm is best to minimize disk I/O and why?

hard

A. Hash join, because building a hash on smaller Table A allows fast matching

B. Sort-merge join, because sorting both tables reduces I/O

C. Cartesian join, because it joins all rows regardless of keys

D. Nested loop join, because it checks all pairs without sorting

Solution

1. Step 1: Consider table sizes and memory

   Table A fits in memory, so building a hash table on it is efficient. Table B is large and unsorted, so sorting it would be expensive.

2. Step 2: Choose join algorithm minimizing disk I/O

   Hash join uses the smaller table in memory to quickly find matches in the larger table, reducing disk reads compared to sorting or nested loops.

3. Final Answer:

   Hash join, because building a hash on smaller Table A allows fast matching -> Option A

4. Quick Check:

   Small table in memory = hash join best [OK]

Hint: Small table fits memory? Use hash join [OK]

Common Mistakes:

• Choosing nested loop for large tables
• Preferring sort-merge without considering sorting cost
• Confusing Cartesian join with normal join