Adventures in Machine Learning

Maximizing SQL Server Performance through Table Partitioning

SQL Server Partitioning: Maximizing Performance and Efficiency

Table partitioning is an essential technique for managing large tables in databases. It works by dividing a table into smaller, more manageable parts that are distributed across different disks or filegroups.

Partitioning enables organizations to improve their databases’ performance by reducing query times, reducing the maintenance workload, and optimizing storage space usage.

Benefits of Table Partitioning

Large tables can take up much space in a database, leading to slower queries and reduced performance. Partitioning tables into smaller manageable parts can help address these issues.

Another benefit of table partitioning is its ability to improve backup and restoration times. Administrators can quickly backup or restore a specific partition, reducing the time it takes to restore the entire table.

Partitioning tables also simplifies maintenance tasks such as data archiving, index maintenance, and data compression.

Creating File Groups

To create partitioned tables in SQL Server, you must first create file groups, which are collections of physical disks or folders where the database files are stored.

You can create two types of file groups: primary and secondary.

A primary file group contains primary data files, while a secondary file group contains only secondary data files.

Creating a Partition Function

A partition function is a database schema object that specifies the partitioning strategy for a table. It determines how data is divided into partitions based on the partitioning column.

To create a partition function in SQL Server, you need to specify the name and data type of the partitioning column, the range boundaries, and the number of partitions.

Creating a Partition Scheme

A partition scheme maps the partition function to a filegroup, specifying which filegroup stores which partition.

To create a partition scheme, you must specify the scheme name, the partition function name, and the filegroup name for each partition.

Creating a Partitioned Table

A partitioned table is a table that is divided into smaller parts based on the partition function. Each partition of a partitioned table is stored in a separate filegroup.

To create a partitioned table, you must specify the name, column name, partition scheme, and clustered index. The partition column is the column used to divide the table into partitions.


Partitioning tables is a vital technique for managing large databases more efficiently. By dividing tables into smaller partitions and storing them in different filegroups, it is easier to manage, query, and perform maintenance tasks.

The ability to back up and restore specific partitions reduces the time it takes to recover data in the event of data loss. By implementing partitioning techniques, organizations can improve database performance and maximize resources’ efficiency.

Example Demonstration: Partitioning Tables to Improve Database Performance

To demonstrate the benefits of partitioning tables, let’s consider an example using the order_reports table. This table contains data on all orders placed by customers on an online marketplace.

Query to Retrieve Order Data

Suppose we want to retrieve data on all orders placed in the third quarter of the year. A simple query to achieve this would be:


FROM order_reports

WHERE order_date BETWEEN ‘2021-07-01’ AND ‘2021-09-30’;

Partitioning the Order_Reports Table

If the order_reports table contains millions of records, retrieving data using the above query can take a while, resulting in slow query performance. To improve the query’s performance, we can partition the order_reports table.

Suppose we decide to partition the order_reports table by order_date. We will partition the table into four partitions, each comprising orders placed in a specific quarter of the year.

To partition the order_reports table:

1. Create the filegroups that will store the table partitions.

CREATE DATABASE orders_database;

ALTER DATABASE orders_database

ADD FILEGROUP data_2021q1;

ALTER DATABASE orders_database

ADD FILEGROUP data_2021q2;

ALTER DATABASE orders_database

ADD FILEGROUP data_2021q3;

ALTER DATABASE orders_database

ADD FILEGROUP data_2021q4;

2. Create the partition function.



VALUES (‘2021-01-01’, ‘2021-04-01’, ‘2021-07-01’, ‘2021-10-01’);

3. Create the partition scheme that maps the filegroups to the partition function.


AS PARTITION pf_order_reports

TO (data_2021q1, data_2021q2, data_2021q3, data_2021q4);

4. Create the partitioned order_reports table and cluster it on the partitioning column.

CREATE TABLE order_reports



order_date DATE,

customer_name VARCHAR(50),

product_name VARCHAR(50),

product_price NUMERIC(10,2)


ON ps_order_reports (order_date);

Inserting Data into the Partitioned Table

We can then insert all the order data into the partitioned table as follows:

INSERT INTO order_reports (order_id, order_date, customer_name, product_name, product_price)


(1, ‘2021-02-01’, ‘John Doe’, ‘Product A’, 100.00),

(2, ‘2021-03-15’, ‘Jane Doe’, ‘Product B’, 150.00),

(3, ‘2021-08-01’, ‘Bob Smith’, ‘Product C’, 200.00),

(4, ‘2021-09-30’, ‘Mary Brown’, ‘Product D’, 50.00);

Checking Rows of Each Partition

To check the rows of each partition, we can run the following query:

SELECT partition_number, rows

FROM sys.partitions

WHERE OBJECT_NAME(object_id)=’order_reports’;

The query result should show four partitions, each with the number of rows contained in them.

Partition_Number Rows

1 1

2 1

3 2

4 0

We can see that the data we inserted was correctly positioned in the appropriate partition.

Retrieving Order Data from the Partitioned Table

Now suppose we want to retrieve data on all orders placed in the third quarter of the year. We can do so using the following query:


FROM order_reports

WHERE order_date BETWEEN ‘2021-07-01’ AND ‘2021-09-30’;

Query performance should be much faster than it would be without partitioning.


In this example demonstration, we explored partitioning a table to improve query performance and optimize resource usage in SQL Server. By partitioning the order_reports table by order_date, inserting data, and using a simple query, we managed to retrieve relevant data much faster than would have been possible without partitioning.

Employing partitioning techniques can help any organization optimize database performance and unlock the full potential of their data. In conclusion, partitioning tables in SQL Server is a crucial technique that can improve database performance, reduce query times, and optimize storage space usage.

By dividing tables into manageable partitions and storing them in separate filegroups, organizations can better manage, query, and perform maintenance tasks. Partitioning tables also simplifies backup and restoration procedures and reduces maintenance workloads.

In summary, partitioning techniques can help organizations unlock the full potential of their data and optimize their database performance. Remember to consider partitioning tables to better manage large data sets and optimize performance.

Popular Posts