CREATE SEQUENCE

Defines a new sequence generator.

Synopsis

CREATE [TEMPORARY | TEMP] SEQUENCE <name>
       [INCREMENT [BY] <value>]
       [MINVALUE <minvalue> | NO MINVALUE]
       [MAXVALUE <maxvalue> | NO MAXVALUE]
       [START [ WITH ] <start>]
       [CACHE <cache>]
       [[NO] CYCLE]
       [OWNED BY { <table>.<column> | NONE }]

Description

CREATE SEQUENCE creates a new sequence number generator. This involves creating and initializing a new special single-row table. The generator will be owned by the user issuing the command.

If a schema name is given, then the sequence is created in the specified schema. Otherwise it is created in the current schema. Temporary sequences exist in a special schema, so a schema name may not be given when creating a temporary sequence. The sequence name must be distinct from the name of any other sequence, table, or view in the same schema.

After a sequence is created, you use the nextval() function to operate on the sequence. For example, to insert a row into a table that gets the next value of a sequence:

INSERT INTO distributors VALUES (nextval('myserial'), 'acme');

You can also use the function setval() to operate on a sequence, but only for queries that do not operate on distributed data. For example, the following query is allowed because it resets the sequence counter value for the sequence generator process on the master:

SELECT setval('myserial', 201);

But the following query will be rejected in HAWQ because it operates on distributed data:

INSERT INTO product VALUES (setval('myserial', 201), 'gizmo');

In a regular (non-distributed) database, functions that operate on the sequence go to the local sequence table to get values as they are needed. In HAWQ, however, keep in mind that each segment is its own distinct database process. Therefore the segments need a single point of truth to go for sequence values so that all segments get incremented correctly and the sequence moves forward in the right order. A sequence server process runs on the master and is the point-of-truth for a sequence in a HAWQ distributed database. Segments get sequence values at runtime from the master.

Because of this distributed sequence design, there are some limitations on the functions that operate on a sequence in HAWQ:

• lastval() and currval() functions are not supported.
• setval() can only be used to set the value of the sequence generator on the master, it cannot be used in subqueries to update records on distributed table data.
• nextval() sometimes grabs a block of values from the master for a segment to use, depending on the query. So values may sometimes be skipped in the sequence if all of the block turns out not to be needed at the segment level. Note that a regular PostgreSQL database does this too, so this is not something unique to HAWQ.

Although you cannot update a sequence directly, you can use a query like:

SELECT * FROM <sequence_name>;

to examine the parameters and current state of a sequence. In particular, the last_value field of the sequence shows the last value allocated by any session.

Parameters

TEMPORARY | TEMP

If specified, the sequence object is created only for this session, and is automatically dropped on session exit. Existing permanent sequences with the same name are not visible (in this session) while the temporary sequence exists, unless they are referenced with schema-qualified names.

CREATE SEQUENCE myseq START 101;

INSERT INTO distributors VALUES (nextval('myseq'), 'acme');

SELECT setval('myseq', 201);

INSERT INTO product VALUES (setval('myseq', 201), 'gizmo');