1. To create a database for your own use, create a user, grant authorization, and log in as instructed in Managing User Permissions. Then, use the CREATE DATABASE statement to create the database. Before database creation, make sure that the logged-in user has CRAETEROLE permissions. For more information on permissions, see Managing User Permissions. A sample database can be created as follows:
CREATE DATABASE test;
All databases can be listed via \\l.
2. A database is usually created based on a template database, and the default template is empty. Any objects in the template will be in the newly created database. You can also specify a template for creation. For example, create a table in the test database first by using the following statement.
create table ttable (age int, id int);
Then, create the test2 database with test as the template.
CREATE DATABASE test2 TEMPLATE test;
Switch to test2, and you can see that ttable also exists in test2. Therefore, avoid creating any objects in template1, as those in template1 will also exist in databases created based on it. You can view all the tables in the selected database with \\d.
3. You can list all the databases with \\l.
4. You can delete the database with DROP DATABASE. When you perform the deletion operation, make sure that the logged-in user is a superuser or a general user with the permission of database deletion. Note that the database can be deleted only if the number of its connections is 0; for example:
You can see that when you select test2, the number of connections to test2 must be greater than or equal to 1. You can only delete test2 after switching to test.
Schema Creation and Management
In CDWPG, a schema is a logical concept for a more detailed division of the database space. Each database has a schema named public when it is created. Tables with the same name cannot be created in the same database unless they are in different schemas. The database system identifies a table in the form of database.schema.table. In addition, schemas with the same name can be created in different databases.
1. Create a schema.
CREATE SCHEMA testschema;
2. Specify the schema to create objects.
When creating tables, functions, or other objects, you can add schema prefixes to indicate that they are to be created in different schemas; if no prefixes are added, objects will be created in the public schema by default; for example:
CREATE TABLE testschema.test;
3. Set the priority of the schema.
ALTER DATABASE testset search_path to testshcema,public;
This statement sets the testschema of the test database to public mode and indicates that testschema has the highest priority. If no schema prefixes are added, testschema will be matched first.
4. Switch the schema.
SET search_path TO public;
If the current schema is testschema, you can switch to public with this statement.
5. Delete the schema.
DROP SCHEMA testschema;
Table Creation and Management
1. Set table and column constraints.
CHECK constraint, which specifies that a data column must satisfy a certain expression, such as:
CREATE TABLE products (product_no int, name text, price int CHECK(price >0));
NOT NULL constraint, which specifies that a data column cannot be empty, such as:
CREATE TABLE products (product_no int NOT NULL, name text NOT NULL, price int CHECK(price >0));
2. Set the data distribution policy.
A distributed database warehouse delivers the best performance when the amount of data stored in each node is the same. If data distribution is unbalanced, nodes with more data will spend more time completing queries, which compromises the overall query performance.
Hash distribution: You can use the DISTRIBUTED BY syntax to specify hash distribution when creating a table. This policy combines all the keys specified as hash distribution and determines the result of data distribution by the hash algorithm. The statement is as follows:
CREATE TABLE test(id int, age int) DISTRIBUTED BY (id);
Random distribution: You can use the DISTRIBUTED RANDOMLY syntax to specify random distribution when creating a table. As the name suggests, this policy determines data distribution randomly. The statement is as follows:
CREATE TABLE test(id int, age int) DISTRIBUTED RANDOMLY;
Here, you must specify one of the data distribution policies for columns with PRIMARY KEY or UNIQUE. For other columns, the first column will be used as the reference for data distribution by default, and the default data distribution policy will be hash distribution.
View Creation and Management
A view is a logical concept. Unlike a table, it has no actual corresponding data structures on the disk.
1. Create a view.
CREATE VIEW testview AS SELECT * FROM ttable where age=28;
Create the testview view with all rows in ttable that satisfy the condition "age=28".
Yes
No
Was this page helpful?