Using Functions and Operators

Description of user-defined and built-in functions and operators in Greenplum Database.

• Using Functions in Greenplum Database
• User-Defined Functions
• Built-in Functions and Operators
• Window Functions
• Advanced Aggregate Functions

Parent topic: Querying Data

Using Functions in Greenplum Database

When you invoke a function in Greenplum Database, function attributes control the execution of the function. The volatility attributes (IMMUTABLE, STABLE, VOLATILE) and the EXECUTE ON attributes control two different aspects of function execution. In general, volatility indicates when the function is executed, and EXECUTE ON indicates where it is executed. The volatility attributes are PostgreSQL based attributes, the EXECUTE ON attributes are Greenplum Database attributes.

For example, a function defined with the IMMUTABLE attribute can be executed at query planning time, while a function with the VOLATILE attribute must be executed for every row in the query. A function with the EXECUTE ON MASTER attribute executes only on the master instance, and a function with the EXECUTE ON ALL SEGMENTS attribute executes on all primary segment instances (not the master).

These tables summarize what Greenplum Database assumes about function execution based on the attribute.

You can display the function volatility and EXECUTE ON attribute information with the psql \df+ function command.

Refer to the PostgreSQL Function Volatility Categories documentation for additional information about the Greenplum Database function volatility classifications.

For more information about EXECUTE ON attributes, see CREATE FUNCTION.

In Greenplum Database, data is divided up across segments — each segment is a distinct PostgreSQL database. To prevent inconsistent or unexpected results, do not execute functions classified as VOLATILE at the segment level if they contain SQL commands or modify the database in any way. For example, functions such as setval() are not allowed to execute on distributed data in Greenplum Database because they can cause inconsistent data between segment instances.

A function can execute read-only queries on replicated tables (DISTRIBUTED REPLICATED) on the segments, but any SQL command that modifies data must execute on the master instance.

Note: The hidden system columns (ctid, cmin, cmax, xmin, xmax, and gp_segment_id) cannot be referenced in user queries on replicated tables because they have no single, unambiguous value. Greenplum Database returns a column does not exist error for the query.

To ensure data consistency, you can safely use VOLATILE and STABLE functions in statements that are evaluated on and run from the master. For example, the following statements run on the master (statements without a FROM clause):

SELECT setval('myseq', 201);
SELECT foo();

If a statement has a FROM clause containing a distributed table and the function in the FROM clause returns a set of rows, the statement can run on the segments:

SELECT * from foo();

Greenplum Database does not support functions that return a table reference (rangeFuncs) or functions that use the refCursor data type.

Function Volatility and Plan Caching

There is relatively little difference between the STABLE and IMMUTABLE function volatility categories for simple interactive queries that are planned and immediately executed. It does not matter much whether a function is executed once during planning or once during query execution start up. But there is a big difference when you save the plan and reuse it later. If you mislabel a function IMMUTABLE, Greenplum Database may prematurely fold it to a constant during planning, possibly reusing a stale value during subsequent execution of the plan. You may run into this hazard when using PREPAREd statements, or when using languages such as PL/pgSQL that cache plans.

User-Defined Functions

Greenplum Database supports user-defined functions. See Extending SQL in the PostgreSQL documentation for more information.

Use the CREATE FUNCTION statement to register user-defined functions that are used as described in Using Functions in Greenplum Database. By default, user-defined functions are declared as VOLATILE, so if your user-defined function is IMMUTABLE or STABLE, you must specify the correct volatility level when you register your function.

By default, user-defined functions are declared as EXECUTE ON ANY. A function that executes queries to access tables is supported only when the function executes on the master instance, except that a function can execute SELECT commands that access only replicated tables on the segment instances. A function that accesses hash-distributed or randomly distributed tables must be defined with the EXECUTE ON MASTER attribute. Otherwise, the function might return incorrect results when the function is used in a complicated query. Without the attribute, planner optimization might determine it would be beneficial to push the function invocation to segment instances.

When you create user-defined functions, avoid using fatal errors or destructive calls. Greenplum Database may respond to such errors with a sudden shutdown or restart.

In Greenplum Database, the shared library files for user-created functions must reside in the same library path location on every host in the Greenplum Database array (masters, segments, and mirrors).

You can also create and execute anonymous code blocks that are written in a Greenplum Database procedural language such as PL/pgSQL. The anonymous blocks run as transient anonymous functions. For information about creating and executing anonymous blocks, see the DO command.

Built-in Functions and Operators

The following table lists the categories of built-in functions and operators supported by PostgreSQL. All functions and operators are supported in Greenplum Database as in PostgreSQL with the exception of STABLE and VOLATILE functions, which are subject to the restrictions noted in Using Functions in Greenplum Database. See the Functions and Operators section of the PostgreSQL documentation for more information about these built-in functions and operators.

Greenplum Database includes JSON processing functions that manipulate values the json data type. For information about JSON data, see Working with JSON Data.

Window Functions

The following built-in window functions are Greenplum extensions to the PostgreSQL database. All window functions are immutable. For more information about window functions, see Window Expressions.

Advanced Aggregate Functions

The following built-in advanced aggregate functions are Greenplum extensions of the PostgreSQL database. These functions are immutable.

Note: The Greenplum MADlib Extension for Analytics provides additional advanced functions to perform statistical analysis and machine learning with Greenplum Database data. See Greenplum MADlib Extension for Analytics in the Greenplum Database Reference Guide.

SELECT departmzent_id, MEDIAN(salary) 
  FROM employees 
GROUP BY department_id; 

CREATE TABLE mymatrix (myvalue int[]);
INSERT INTO mymatrix 
   VALUES (array[[1,2],[3,4]]);
INSERT INTO mymatrix 
   VALUES (array[[0,1],[1,0]]);
SELECT sum(myvalue) FROM mymatrix;
 sum 
———————-
 {{1,3},{4,4}}

Need help? Visit the Greenplum Database Community