การใช้งานการแบ่งพาร์ทิชันแบบรายการ

Welcome! In this lesson, we'll dive into List Partitioning, a powerful way to organize your data in PostgreSQL.

List partitioning divides a table based on discrete values in a specified column. Think of it as categorizing your data into distinct groups.

You might recall Range Partitioning from a previous lesson, which divides data based on a *range* of values (e.g., dates, IDs).

List partitioning is different: it uses specific, distinct values. For example, you could partition by country_code ('US', 'CA', 'MX') or order_status ('PENDING', 'SHIPPED', 'CANCELLED').

The column you choose to partition by is called the partition key. For list partitioning, this column should have a finite and manageable set of distinct values.

• Good candidates: Country codes, product categories, order statuses.
• Poor candidates: Timestamps, large text fields, unique IDs (too many distinct values).

First, we create the main table, known as the parent table. We specify PARTITION BY LIST and indicate the column that will serve as our partition key.

Let's create a sales table partitioned by country_code:

Once the parent table is ready, we create individual child partitions. Each child partition is a separate table that stores rows for specific values of the partition key.

Here, we create partitions for sales in the 'US' and 'CA':

When you insert data into the parent sales table, PostgreSQL automatically directs each row to the correct child partition based on its country_code value.

Let's add some sales data and then check the individual partitions:

It's good practice to verify that your partitions are correctly linked to the parent table. You can query PostgreSQL's catalog tables to see the inheritance structure.

This query lists child tables for 'sales':

As your business expands or new data categories emerge, you can easily add new partitions to your existing scheme without downtime.

Let's add a partition for Mexico ('MX') sales and insert some data:

What happens if you insert a country_code that isn't explicitly covered by a partition (e.g., 'GB')?

You can create a default partition to catch all values that don't match any other defined list. This prevents errors for unexpected data.

Consider a table orders partitioned by list on the status column, which can be 'NEW', 'PROCESSING', 'SHIPPED', or 'CANCELLED'.

We've covered how to implement list partitioning in PostgreSQL. You learned to:

• Define a parent table with PARTITION BY LIST.
• Create child partitions using FOR VALUES IN.
• Route data automatically upon insertion.
• Add new partitions and handle unspecified values with a DEFAULT partition.

List partitioning is ideal for data with discrete, categorical values, helping organize and query your data more efficiently.