Partitioning in Oracle Database: Because Who Wants to Search an Endless Table?

As data volumes continue to grow, managing large database tables and indexes can become a challenge. This is where partitioning comes in. Partitioning is a feature of database systems that allows you to divide large tables and indexes into smaller, more manageable parts, known as partitions. This can improve the performance and manageability of your database. Aside from performance considerations, maintenance operations, such as backups and index rebuilds, can become easier by allowing them to be performed on smaller subsets of data.

This is achieved by reducing the amount of data that needs to be scanned during query execution. When a query is executed, the database can use the partitioning information to skip over partitions that do not contain the relevant data, instead of having to scan the entire table. This reduces the amount of I/O required to execute the query, which can result in significant performance gains, especially for large tables.

There are several types of partitioning available in Oracle Database, including range partitioning, hash partitioning, list partitioning, and composite partitioning. Each type of partitioning is suited to different use cases and can be used to optimize the performance of your database in different ways. In this blog post we will look range partitioning.

Range partitioning

Here is an example of range-based partitioning in Oracle:

CREATE TABLE books (
  id NUMBER,
  title VARCHAR2(200),
  publication_year NUMBER
)

PARTITION BY RANGE (publication_year) (
  PARTITION p_before_2000 VALUES LESS THAN (2000),
  PARTITION p_2000s VALUES LESS THAN (2010),
  PARTITION p_2010s VALUES LESS THAN (2020),
  PARTITION p_after_2020 VALUES LESS THAN (MAXVALUE)
);

In this example, we have created a table called books that stores book titles, partitioned by the year of publication. We have defined four partitions, p_before_2000, p_2000s, p_2010s, and p_after_2020.

Now, when we insert data into the books table, it will automatically be placed in the appropriate partition based on the year of publication:

INSERT INTO books (id, title, publication_year)
  VALUES (1, 'Nineteen Eighty-Four', 1949);

This book will be inserted into partition p_before_2000, as the year of publication is before 2000. The following book will be placed into partition p_2000s:

INSERT INTO books (id, title, publication_year)
  VALUES (2, 'The Hunger Games', 2008);

When we query the books table, the database will only access the partitions that contain the data we need. For example, if we want to retrieve data for books published in 2015 and 2016, the database will only access partition p_2010s.

SELECT * FROM books WHERE publication_year>=2015 AND publication_year<=2016:

However, you should be aware that while partitioning can improve query performance for some types of queries, it can also negatively impact query performance for others, especially if the partitioning scheme does not align well with the query patterns. Therefore, you should tailor the partitioning to your needs and check if it brings the desired effect.