Records

Updated on 2025-03-13 GMT+08:00

View PDF

Record Variables

Perform the following operations to create a record variable:

Define a record type and use this type to declare a variable.

Syntax

For the syntax of the record type, see Figure 1.

Figure 1 Syntax of the record type
Click to enlarge

The above syntax diagram is explained as follows:

type_name: specifies the record type.
field_name: specifies member variables of this type.
datatype: specifies data types for the member variables.
expression: specifies the expression for setting a default value.

NOTE:

In GaussDB:

When assigning values to record variables, you can:
- Declare a record type and define member variables of this type when you declare a function or stored procedure.
- Assign the value of a record variable to another record variable.
- Use SELECT INTO or FETCH to assign values to a record type.
- Assign the NULL value to a record variable.
The INSERT and UPDATE statements cannot use a record variable to insert or update data.
The constructor of the record type cannot be used as the default value of a function or stored procedure parameter.
When package_name.record_type is used to declare a record variable, the default value of the record type becomes invalid. Therefore, you are advised not to use package_name.record_type to declare a record variable when the record type has a default value.
Just like a variable, a record column of the compound type does not have a default value in the declaration.
If a member has a record type, the default value of the inner record type is not transferred to the outer record type.
If enable_recordtype_check_strict is set to on and the member is of the record type and a column of the record type has the not null or default attribute, an error is reported during stored procedure or package compilation.
If enable_recordtype_check_strict is set to on, a package contains the record type, and a column of the record type has the not null or default attribute, an error is reported during creation or compilation.
If enable_recordtype_check_strict is set to on, the not null function of the record type in the stored procedure takes effect.
The data type can be the record type, array type, or collection type defined in a stored procedure (anonymous blocks are not supported).
After the GUC parameter behavior_compat_options is set to 'proc_outparam_override';:
- Functions that contain the out output parameter and returns data of the record type can be called using SELECT and CALL in external SQL statements, and can be called using PERFORM and expressions in stored procedures.
- If the function returns data of the undefined record type, at least one out parameter is required. If the function returns data of the defined record type, the out parameter is optional. For details, see the example.
- Functions that contain the out output parameter and return data of the composite or record type, ensure that the expected return type of the function is the same as the actual return type.

Examples

    
     
       
       -- Create a table and insert data into the table.
gaussdb=# CREATE TABLE emp_rec(
gaussdb(#     empno numeric(4,0) not null,
gaussdb(#     ename varchar(10)
gaussdb(# );
CREATE TABLE
gaussdb=# INSERT INTO emp_rec VALUES(111, 'aaa'), (222, 'bbb'), (333, 'ccc');
INSERT 0 3
-- The table is defined as follows:
gaussdb=# \d emp_rec
           Table "public.emp_rec"
 Column |         Type          | Modifiers 
--------+-----------------------+-----------
 empno  | numeric(4,0)          | not null
 ename  | character varying(10) |

-- Perform record operations in the function.
gaussdb=# CREATE OR REPLACE FUNCTION regress_record(p_w VARCHAR2) RETURNS VARCHAR2 AS $$
gaussdb$# DECLARE
gaussdb$#     -- Declare a record type.
gaussdb$#     type rec_type is record (name  varchar2(100), epno int);
gaussdb$#     employer rec_type;
gaussdb$#     -- Use %type to declare the record type.
gaussdb$#     type rec_type1 is record (name  emp_rec.ename%type, epno int :=10);
gaussdb$#     employer1 rec_type1;
gaussdb$#     -- Declare a record type with the default value.
gaussdb$#     type rec_type2 is record (
gaussdb$#         name varchar2 not null := 'SCOTT',
gaussdb$#         epno int not null :=10
gaussdb$#     );
gaussdb$#     employer2 rec_type2;
gaussdb$#     CURSOR C1 IS  select ename,empno from emp_rec order by 1 limit 1;
gaussdb$# BEGIN
gaussdb$#     -- Assign a value to a member record variable.
gaussdb$#     employer.name := 'WARD';
gaussdb$#     employer.epno = 18;
gaussdb$#     raise info 'employer name: % , epno:%', employer.name, employer.epno;
gaussdb$# 
gaussdb$#     -- Assign the value of a record variable to another variable.
gaussdb$#     employer1 := employer;
gaussdb$#     raise info 'employer1 name: % , epno: %',employer1.name, employer1.epno;
gaussdb$# 
gaussdb$#     -- Assign the NULL value to a record variable.
gaussdb$#     employer1 := NULL;
gaussdb$#     raise info 'employer1 name: % , epno: %',employer1.name, employer1.epno;
gaussdb$# 
gaussdb$#     -- Obtain the default value of a record variable.
gaussdb$#     raise info 'employer2 name: % ,epno: %', employer2.name, employer2.epno;
gaussdb$# 
gaussdb$#     -- Use a record variable in the FOR loop.
gaussdb$#     for employer in select ename,empno from emp_rec order by 1 limit 1 loop 
gaussdb$#         raise info 'employer name: % , epno: %', employer.name, employer.epno;
gaussdb$#     END loop;
gaussdb$# 
gaussdb$#     -- Use a record variable in the SELECT INTO statement.
gaussdb$#     select ename,empno  into employer2 from emp_rec order by 1 limit 1;
gaussdb$#     raise info 'employer name: % , epno: %', employer2.name, employer2.epno;
gaussdb$# 
gaussdb$#     -- Use a record variable in a cursor.
gaussdb$#     OPEN C1;
gaussdb$#     FETCH C1 INTO employer2;
gaussdb$#     raise info 'employer name: % , epno: %', employer2.name, employer2.epno;
gaussdb$#     CLOSE C1;
gaussdb$#     RETURN employer.name;
gaussdb$# END; $$
gaussdb-# LANGUAGE plpgsql;
CREATE FUNCTION

-- Call this function.
gaussdb=# CALL regress_record('abc');
INFO:  employer name: WARD , epno:18
INFO:  employer1 name: WARD , epno: 18
INFO:  employer1 name: <NULL> , epno: <NULL>
INFO:  employer2 name: SCOTT ,epno: 10
INFO:  employer name: aaa , epno: 111
INFO:  employer name: aaa , epno: 111
INFO:  employer name: aaa , epno: 111
 regress_record 
----------------
 aaa
(1 row)
-- Drop the function.
gaussdb=# DROP FUNCTION regress_record;
DROP FUNCTION
-- Drop the table.
gaussdb=# DROP TABLE emp_rec;
DROP TABLE

-- If the compatibility parameter proc_outparam_override is enabled, the data of the defined record type is returned. The out parameter is not required for the function.
gaussdb=# CREATE TYPE rec_type IS (c1 int, c2 int);
CREATE TYPE
gaussdb=# SET behavior_compat_options = 'proc_outparam_override';
SET
gaussdb=# CREATE OR REPLACE FUNCTION func(a in int) RETURN rec_type IS
gaussdb$#     r rec_type;
gaussdb$# BEGIN
gaussdb$#     r.c1:=1;
gaussdb$#     r.c2:=1;
gaussdb$#     return r;
gaussdb$# END;
gaussdb$# /
CREATE FUNCTION
gaussdb=# CALL func(0);
 c1 | c2 
----+----
  1 |  1
(1 row)
gaussdb=# DROP FUNCTION func;
DROP FUNCTION
gaussdb=# drop type rec_type;
DROP TYPE
-- If the compatibility parameter proc_outparam_override is enabled, the data of the undefined record type is returned and at least one out parameter is required for the function
gaussdb=# SET behavior_compat_options = 'proc_outparam_override';
SET
gaussdb=# CREATE OR REPLACE FUNCTION func(a out int) RETURN record IS
gaussdb$#     type rc is record(c1 int, c2 int);
gaussdb$#     r rc;
gaussdb$# BEGIN
gaussdb$#     r.c1 := 1;
gaussdb$#     r.c2 := 1;
gaussdb$#     a := 999;
gaussdb$#     return r;
gaussdb$# END;
gaussdb$# /
CREATE FUNCTION
gaussdb=# CALL func(1);
 func  |  a  
-------+-----
 (1,1) | 999
(1 row)
gaussdb=# DROP FUNCTION func;
DROP FUNCTION

      

    
   

Parent topic: Arrays, Collections, and Records

Previous topic: Collection Functions

Next topic: DECLARE Syntax

Feedback

Was this page helpful?

Helpful Not helpful

Provide feedback

Thank you very much for your feedback. We will continue working to improve the documentation.See the reply and handling status in My Cloud VOC.

The system is busy. Please try again later.

Which of the following issues have you encountered?

Content is inconsistent with the product UI

Unclear descriptions

Lack of examples or code

Incorrect steps

Can't find what I need

Lack of best practices

Feedback (optional)

0/500

Select at least one type of issue, and enter your comments or suggestions.

Enter a maximum of 500 characters.

Submit Cancel

For any further questions, feel free to contact us through the chatbot.

Chatbot