''

Assignment 4: Indexing and Query Optimization in DBMS

Page 1: Introduction to Indexing in DBMS

Indexing is a crucial technique used in database management systems to optimize the speed
and efficiency of data retrieval. When a database contains millions of records, searching for data
without indexing would mean scanning the entire dataset, which is very time-consuming and
inefficient.

An index in a database is similar to an index in a book; it allows the database engine to find the
data quickly without searching every row in a table. Indexes are special data structures that store
key values and pointers to the actual data rows in the table.

The importance of indexing cannot be overstated as it significantly reduces the number of disk I/
O operations and speeds up query performance, especially for read-heavy applications.

Page 2: Types of Indexes

There are several types of indexes in DBMS, each suited for different use cases:

1. Primary Index:
- Built on the primary key of a table.
- The index is unique and sorted.
- One primary index per table.

2. Secondary Index:
- Created on non-primary key columns.
- Can have multiple secondary indexes per table.
- Useful for searching based on non-key attributes.

3. Clustered Index:
- Determines the physical order of data in the table.
- Only one clustered index per table.
- Data rows are stored in sorted order.

4. Non-Clustered Index:
- Separate structure from the data table.
- Contains pointers to the actual data rows.
- Multiple non-clustered indexes allowed.

5. Composite Index:
- Index on multiple columns.
- Helps in queries filtering on several attributes.

Page 3: Data Structures for Indexing

The choice of data structure affects the performance and efficiency of indexing.

1. B-Tree Index:
- A balanced tree data structure.
- Each node contains multiple keys and pointers.
- Supports efficient search, insert, delete in O(log n) time.
- Commonly used in databases.

2. B+ Tree Index:
- A variant of B-tree.
- All data records stored at leaf nodes.
- Leaf nodes linked sequentially for efficient range queries.

3. Hash Index:
- Uses a hash function to map keys to buckets.
- Extremely fast for equality searches.
- Not suitable for range queries.

Page 4: Dense and Sparse Indexes

- Dense Index: Contains index entries for every search key value in the database.
- Sparse Index: Contains entries for only some records, usually one per data block.

Dense indexes provide faster access but require more storage space and maintenance.

Page 5: Multi-level Indexing

When the index itself grows large, searching it can become slow. Multi-level indexing solves this
by creating an index on the index.

- First-level index points to blocks of second-level indexes.

- Second-level indexes point to data blocks.
- This hierarchy reduces the number of disk reads.

Page 6: Index Maintenance and Overhead

Indexes improve read performance but come with overhead:

- Insertion, deletion, and updates require maintaining indexes.

- Indexes consume additional storage space.
- Too many indexes can degrade write performance.

Therefore, indexing strategy must balance query speed and maintenance overhead.

Page 7: Introduction to Query Optimization

Query Optimization is the process of choosing the most efficient way to execute a given query by
considering possible query plans. Query optimizers aim to minimize resource use such as CPU
time, memory, and disk I/O.

The process typically involves:

- Parsing the query.

- Translating it into a relational algebra expression.
- Generating possible execution plans.
- Estimating costs.
- Selecting the best plan.

Page 8: Query Execution Plans

An execution plan is a sequence of operations the database engine will perform to answer the
query.

Operations include:

- Scans (table scan, index scan).

- Joins (nested loop, merge join, hash join).
- Sorting and aggregation.

Understanding execution plans helps optimize slow queries.

Page 9: Join Algorithms

Joins are often the most expensive operations in queries.

1. Nested Loop Join:
- For each tuple in outer relation, search inner relation.
- Simple but costly for large datasets.

2. Merge Join:
- Requires sorted inputs.
- Efficient for large, sorted datasets.

3. Hash Join:
- Uses a hash table to match tuples.
- Good for large, unsorted data.

Page 10: Cost Estimation and Statistics

Optimizers estimate the cost of query plans using statistics like:

- Number of rows in tables.

- Data distribution.
- Available indexes.
- Selectivity of predicates.

Better statistics lead to better optimization.

Page 11: Heuristics and Rule-Based Optimization

Besides cost-based methods, query optimizers use heuristics such as:

- Push selections and projections down the query tree.

- Join smaller tables first.
- Use indexes where possible.

These rules simplify optimization.

Page 12: Challenges in Query Optimization

- Complex queries with multiple joins.

- Dynamic data distributions.
- Accurate statistics collection.
- Balancing optimization time with execution time.
Page 13: Indexing and Optimization in Real Systems

Modern DBMS use advanced indexing (like bitmap indexes, full-text indexes) and sophisticated
optimizers.

Examples:

- Oracle uses cost-based optimization.

- SQL Server provides execution plans and index tuning advisors.
- PostgreSQL has a flexible optimizer and supports multiple index types.

Page 14: Conclusion

Indexing and query optimization are foundational for database performance. Effective indexing
strategies combined with powerful optimizers ensure fast and reliable data retrieval.
Understanding these concepts helps in designing databases and writing queries that scale
efficiently.

''