Introduction
Cursor is a temporary memory area (context area) where Oracle executes SQL statements. Oracle associates every SELECT
statement with a cursor to hold the query information in this context area.
Types of Cursor
There are two types of cursors - implicit and explicit cursor.
Implicit cursor: Oracle automatically (implicit) controls or processes the information of SQL statement executed. In this process, the user is unaware of implicit cursor. Oracle automatically performs the OPEN, FETCH,
and CLOSE
operations.
Explicit cursor: Explicit cursor is used for the query that returns more than one row of data. These cursors are explicitly declared in the DECLARE
section of the PL/SQL block. This declaration allows to sequentially process each row of data as the cursor returns it. In explicit cursor DECLARE,OPEN,FETCH,
and CLOSE
operations are done by the programmer.
The process of working with an explicit cursor:
- Declare: The cursor is initialised into temporary memory area.
- Open: The cursor is opened which is declared, and the temporary memory area is allotted.
- Fetch: Cursor which is declared and opened can now retrieve rows from data.
- Close: The
CLOSE
statement disables the cursor, and releases the temporary memory area.
Cursor Attributes
These attributes return useful information about the execution of a SQL statement. Cursor attributes can be used in procedural statements but not in SQL statements.
CURSOR ATTRIBUTE | SYNTAX | DESCRIPTION |
---|
%NOTFOUND
|
cursor_name%NOTFOUND
| %NOTFOUND returns TRUE if last fetch did not return a row, Else FALSE if last fetch returns row. |
%FOUND
|
cursor_name%FOUND
| %FOUND returns TRUE if the cursor is open, fetches the row till the last fetch. FALSE if last fetch did not return any row. |
%ROWCOUNT
|
cursor_name%ROWCOUNT
| %ROWCOUNT keeps track of fetched rows from cursor until it is closed. |
%ISOPEN
|
cursor_name%ISOPEN
| %ISOPEN returns TRUE if its cursor or cursor variable is open, otherwise, %ISOPEN returns FALSE . |
The general syntax for creating a cursor is as follows :
CURSOR cursor_name IS select_statement;
- cursor_name : name for the cursor
- select_statement :
select
query which returns multiple rows
As we have gone through an introductory part , let's start exploring cursor with examples to understand more in depth.
NOTE : In programming language, standardisation is the most important part. When you declare a cursor in declaration part, always start with initial c_cursor_name
, it will always be clear that name refers to cursor.
Before we start with Explicit cursor, let us understand...
Record Types
Record is composite data structure, it is group of data items similar to the row of database table, each item with its own name and datatype. Assume the record is a variable which holds a table row or some columns of a table. If you want to learn more about RECORD
, you can refer to PL/SQL User's Guide and Reference chapter 5.
PL/SQL supports only three types of records - table based, cursor based and programmer defined.
- A table based record is one whose entire structure is similar to columns of a table.
- A cursor based record is one whose structure is similar to items (elements) of a pre-defined cursor.
- A cursor and record datatype created in declaration part of PL/SQL Block having the same variables with same datatype in the same order is called programmer defined record.
NOTE: To create a table based and cursor based record, always prefer to use %ROWTYPE
attribute. The %ROWTYPE
attribute provides a record type that represents a row in a database table. The record can store an entire row of data selected from the table or fetched from a cursor or cursor variable. To learn more about %ROWTYPE
, refer to PL/SQL User's Guide and Reference chapter 2.
Example 1. Table based record
DECLARE
vr_emp employees%ROWTYPE;
BEGIN
SELECT *
INTO vr_emp
FROM employees
WHERE employee_id = 100;
DBMS_OUTPUT.PUT_LINE('Employee Details : '||vr_emp.employee_id
||' '||vr_emp.first_name||' '||vr_emp.last_name||' '||vr_emp.salary);
END;
Employee Details : 100 Steven King 24000
Example 2. Cursor based record
DECLARE
CURSOR c_emp_detail IS
SELECT employee_id,first_name,last_name,salary
FROM employees;
rec_emp_detail c_emp_detail%ROWTYPE;
BEGIN
OPEN c_emp_detail;
LOOP
FETCH c_emp_detail INTO rec_emp_detail;
EXIT WHEN c_emp_detail%NOTFOUND;
DBMS_OUTPUT.PUT_LINE('Employees Details : '||' '||rec_emp_detail.employee_id
||' '||rec_emp_detail.first_name||' '||rec_emp_detail.last_name);
END LOOP;
DBMS_OUTPUT.PUT_LINE('Total number of rows : '||c_emp_detail%ROWCOUNT);
CLOSE c_emp_detail;
END;
Employees Details : 198 Donald OConnell
Employees Details : 199 Douglas Grant
Employees Details : 200 Jennifer Whalen
-----------------------------
Employees Details : 196 Alana Walsh
Employees Details : 197 Kevin Feeney
Total number of rows : 107
Example 3. Programmer defined record
DECLARE
CURSOR c_emp_detail IS
SELECT employee_id,first_name,last_name,salary
FROM employees;
TYPE type_rectype IS RECORD
(emp_id employees.employee_id%TYPE,
f_name employees.first_name%TYPE,
l_name employees.last_name%TYPE,
s_salary employees.salary%TYPE
);
rec_type type_rectype;
BEGIN
OPEN c_emp_detail;
LOOP
FETCH c_emp_detail INTO rec_type;
EXIT WHEN c_emp_detail%NOTFOUND;
DBMS_OUTPUT.PUT_LINE('Employees Details : '||' '||rec_type.emp_id
||' '||rec_type.f_name||' '||rec_type.l_name||' '||rec_type.s_salary);
END LOOP;
DBMS_OUTPUT.PUT_LINE('Total number of Employees : '||c_emp_detail%ROWCOUNT);
CLOSE c_emp_detail;
END;
Employees Details : 198 Donald OConnell 2600
Employees Details : 199 Douglas Grant 2600
Employees Details : 200 Jennifer Whalen 4400
-----------------------------
Employees Details : 196 Alana Walsh 3100
Employees Details : 197 Kevin Feeney 3000
Total number of Employees : 107
Cursors in Nested Loops
Before going through Example 4, please make sure you have a better understanding of Nested loop concept, because Cursors in Nested Loops concept is absolutely the same as Nested Loops. To learn more about Loops, refer to PL/SQL User's Guide and Reference chapter 4.
Example 4. Cursors in Nested Loops
DECLARE
CURSOR c_dept IS
SELECT *
FROM departments
WHERE manager_id IS NOT NULL
ORDER BY department_name;
r_dept c_dept%ROWTYPE;
CURSOR c_emp (c_dept_no departments.department_id%TYPE) IS
SELECT *
FROM employees
WHERE department_id = c_dept_no;
r_emp c_emp%ROWTYPE;
BEGIN
OPEN c_dept;
LOOP
FETCH c_dept INTO r_dept;
EXIT WHEN c_dept%NOTFOUND;
DBMS_OUTPUT.PUT_LINE('----------------------------------');
DBMS_OUTPUT.PUT_LINE('Department Name : '||r_dept.department_name);
DBMS_OUTPUT.PUT_LINE('----------------------------------');
OPEN c_emp(r_dept.department_id);
LOOP
FETCH c_emp INTO r_emp;
EXIT WHEN c_emp%NOTFOUND;
DBMS_OUTPUT.PUT_LINE('Employees Details : '||r_emp.employee_id
||' '||r_emp.first_name||' '||r_emp.last_name||' '||r_emp.salary);
END LOOP;
CLOSE c_emp;
END LOOP;
CLOSE c_dept;
END;
----------------------------------
Department Name : Accounting
----------------------------------
Employees Details : 205 Shelley Higgins 12000
Employees Details : 206 William Gietz 8300
-----------------------------
----------------------------------
Department Name : Shipping
----------------------------------
Employees Details : 198 Donald OConnell 2600
Employees Details : 199 Douglas Grant 2600
-----------------------------
The c_emp
cursor specifies the parameter in c_dept_no
. Each time cursor c_emp
is called, it returns only the list of employees of which the department_id
specified by the parameter (c_dept_no
). In Example 4, if we segregate the individual cursor c_emp
, it becomes a parameterised cursor.
Cursor For Loops
In Cursor for loops, the process of opening, fetching and closing is handled implicitly. This makes the programmer code and maintain the blocks easily.
In cursor for loop, before each iteration PL/SQL fetches into implicitly declared record.
The sequence of statements inside the loop is executed once for each row that satisfies the query.
When loop is left, the cursor is automatically closed.
The cursor is closed even if you use an EXIT
or GOTO
statement to leave the loop before all rows are fetched.
Example 5. Implicit cursor for loop
BEGIN
FOR item IN(SELECT department_name,d.department_id,last_name,job_id,salary
FROM departments d JOIN employees e
ON e.department_id = d.department_id
WHERE JOB_ID = 'IT_PROG'
AND salary > 4800)
LOOP
DBMS_OUTPUT.PUT_LINE(item.last_name||' '||item.department_name
||' '||item.department_id||' '||item.job_id||' '||item.salary);
END LOOP;
END;
Hunold IT 60 IT_PROG 9000
Ernst IT 60 IT_PROG 6000
If you need the same query to reference from different parts of the same procedure, you can declare a cursor with that specific query,and get the results using cursor for loop. For this, I am going to use the same select
query from Example 5.
Example 6. Explicit cursor for loop
DECLARE
CURSOR c_detail IS
SELECT department_name,d.department_id,last_name,job_id,salary
FROM departments d JOIN employees e
ON e.department_id = d.department_id
WHERE JOB_ID = 'IT_PROG'
AND salary > 4800;
BEGIN
FOR item IN c_detail
LOOP
DBMS_OUTPUT.PUT_LINE(item.last_name||' '||item.department_name
||' '||item.department_id||' '||item.job_id||' '||item.salary);
END LOOP;
END;
Hunold IT 60 IT_PROG 9000
Ernst IT 60 IT_PROG 6000
Example 7. Nested Cursors Using Cursor FOR Loops
If you notice, example 4 and example 7 outputs are the same, the difference is example 7. We are using cursor for loop where record variable declares r_dept
and r_emp
, opening fetching and closing is done automatically by each loop iteration until all rows are fetched according to the specific query in cursor.
The variable v_dept_id
is initialized to be the department_id
of the current record of the c_dept
cursor. The c_dept
cursor ties in the c_emp
cursor by means of this variable.
Thus, when the cursor c_emp
is processed, it is retrieving employees who have the department_id
of the current record for the c_dept
cursor.
Each iteration of the c_dept
cursor will execute the DBMS_OUTPUT
only once. The DBMS_OUTPUT
will be executed once for each iteration of the c_emp
cursor loop, producing a line of output for each employee.
DECLARE
v_dept_id departments.department_id%TYPE;
CURSOR c_dept IS
SELECT *
FROM departments
WHERE manager_id IS NOT NULL
ORDER BY department_name;
CURSOR c_emp IS
SELECT *
FROM employees
WHERE department_id = v_dept_id;
BEGIN
FOR r_dept IN c_dept
LOOP
v_dept_id := r_dept.department_id;
DBMS_OUTPUT.PUT_LINE('----------------------------------');
DBMS_OUTPUT.PUT_LINE('Department Name : '||r_dept.department_name);
DBMS_OUTPUT.PUT_LINE('----------------------------------');
FOR r_emp IN c_emp
LOOP
DBMS_OUTPUT.PUT_LINE('Employee Name : '||r_emp.last_name);
END LOOP;
END LOOP;
END;
----------------------------------
Department Name : Accounting
----------------------------------
Employee Name : Higgins
Employee Name : Gietz
----------------------------------
--------------------
----------------------------------
Department Name : Shipping
----------------------------------
Employee Name : OConnell
Employee Name : Grant
--------------------
Example 8. Using Parameters with Nested Cursors For Loops
CREATE OR REPLACE PROCEDURE print_emp_dept(v_lo_id IN locations.location_id%TYPE)
IS
v_flag departments.department_id%TYPE;
CURSOR c_locations IS
SELECT *
FROM locations
WHERE location_id = v_lo_id;
CURSOR c_departments(v_loc_id locations.location_id%TYPE) IS
SELECT l.location_id,department_name,department_id
FROM locations l JOIN departments d
ON l.location_id = d.location_id
WHERE l.location_id = v_loc_id
AND d.manager_id IS NOT NULL;
CURSOR c_employees (v_dept_id departments.department_id%TYPE,
v_loc_id locations.location_id%TYPE) IS
SELECT d.department_id,employee_id,first_name,last_name,salary,job_id,city
FROM locations l JOIN departments d
ON l.location_id = d.location_id
JOIN employees e
ON d.department_id = e.department_id
WHERE d.department_id = v_dept_id
AND l.location_id = v_loc_id;
BEGIN
FOR r_location IN c_locations
LOOP
DBMS_OUTPUT.PUT_LINE
('Location ID : '||r_location.location_id||
' Belong''s to '||r_location.city||' city');
DBMS_OUTPUT.PUT_LINE('In city '||r_location.city||' '||'Departments are ');
FOR r_department IN c_departments(r_location.location_id)
LOOP
DBMS_OUTPUT.PUT_LINE('-----------------------------------------');
DBMS_OUTPUT.PUT_LINE('Department ID :
'||r_department.department_id||' '||'Location ID : '||
r_department.location_id||' '||
'Department Name : '||r_department.department_name);
DBMS_OUTPUT.PUT_LINE('-----------------------------------------');
v_flag := r_department.department_id;
EXIT WHEN v_flag IS NULL;
FOR r_employee IN c_employees
(r_department.department_id,r_location.location_id)
LOOP
DBMS_OUTPUT.PUT_LINE(r_employee.employee_id||'
'||r_employee.first_name||' '||
r_employee.last_name||' work''s in city '||r_employee.city);
END LOOP;
END LOOP;
END LOOP;
IF v_flag IS NULL THEN
DBMS_OUTPUT.PUT_LINE('Not Available');
END IF;
END print_emp_dept;
Call the procedure:
CALL print_emp_dept(3000);
BEGIN
print_emp_dept(1700);
END;
Location ID : 1400 Belong's to Southlake city
In city Southlake Departments are
-----------------------------------------
Department ID :60 Location ID : 1400 Department Name : IT
-----------------------------------------
103 Alexander Hunold work's in city Southlake
104 Bruce Ernst work's in city Southlake
105 David Austin work's in city Southlake
106 Valli Pataballa work's in city Southlake
107 Diana Lorentz work's in city Southlake
The grandparent cursor, c_locations
is declared. It is a collection of locations and takes parameter while call of procedure print_emp_dept
with a location_id
is available in database.
The parent cursor, c_departments
is declared. It takes in the parameter of location_ID
to generate a list of departments that belongs to parameter location_ID
.
The child cursor, c_employees
takes in two parameters, department_id
and location_id
. In this way, it generates a list of employees working with different departments but in the same location_id
.
The grandparent cursor loop begins, and only the city name and location_id
is displayed with DBMS_OUTPUT
. The parent cursor loop begins. It takes the parameter of location_id
from the grandparent cursor where department_id
, location_id
and department_name
is displayed.
The child cursor loop begins. It takes in the parameter of department_id
from the parent cursor and location_id
from the grandparent cursor.The employee_id
and employees name is displayed in which city they work because of the parameter which has been used from grandparent cursor.
In this complete process, the child cursor loop ends first, then parent cursor loop ends, then grandparent cursor loop ends.
Example 9. For UPDATE and WHERE CURRENT CLAUSE
The FOR UPDATE
clause is only used in cursor, when update
or delete
statements are used for database tables. Normally, when programmer executes SE<code>
LECT statement, there is no locking of rows. The main aim to use FOR UPDATE
clause is to lock rows when performing update
or delete
statements inside the cursor, and restrict other users to perform any updation in particular database tables. Once the updation is done inside the cursor, then COMMIT
or ROLLBACK
placed inside the execution block releases the lock. Now FOR UPDATE
clause with particular Column name i.e.. FOR UPDATE
salary will only lock salary column according to select
statement even if there is join
condition, else FOR UPDATE
clause will lock entire row of that particular table.
NOWAIT
is an optional keyword, if the rows are already locked by another programmer, then control is immediately returned to programmer so that meanwhile he can do other work before trying again. If you omit the keyword, then the wait may be a long period of time.
The WHERE CURRENT OF
clause can be used only if FOR UPDATE
clause is used in cursor.
The WHERE
current of clause only references the cursor which fetches the latest row.
The WHERE CURRENT OF CLAUSE
is useful to eliminate the where condition in update
clause.
First, I have created a demo table, because I don't want to change data in the actual table of database.
create table emp1 as select * from employees;
create table dept1 as select * from departments;
DECLARE
CURSOR c_sal_update IS
SELECT employee_id,first_name,last_name,job_id,department_name,e.department_id,salary
FROM dept1 d , emp1 e
WHERE e.department_id = 80
FOR UPDATE OF salary NOWAIT;
rec_sal c_sal_update%ROWTYPE;
BEGIN
OPEN c_sal_update;
LOOP
FETCH c_sal_update INTO rec_sal;
EXIT WHEN c_sal_update%NOTFOUND;
IF rec_sal.job_id = 'SA_MAN' THEN
UPDATE emp1
SET salary = rec_sal.salary + 1000
WHERE CURRENT OF c_sal_update;
END IF;
END LOOP;
COMMIT;
CLOSE c_sal_update;
END;
OUTPUT
SELECT employee_id,first_name,last_name,job_id,department_id,salary
FROM emp1
WHERE job_id = 'SA_MAN'
AND department_id = 80;
EMPLOYEE_ID FIRST_NAME LAST_NAME JOB_ID DEPARTMENT_ID SALARY
145 John Russell SA_MAN 80 15000
146 Karen Partners SA_MAN 80 14500
147 Alberto Errazuriz SA_MAN 80 13000
148 Gerald Cambrault SA_MAN 80 12000
149 Eleni Zlotkey SA_MAN 80 11500
Example 10. Cursor attributes with explicit cursor
DECLARE
CURSOR c_high_sal IS
SELECT *
FROM (SELECT employee_id,first_name,last_name,salary
FROM employees ORDER BY salary DESC)
WHERE ROWNUM < 11;
high_sal c_high_sal%ROWTYPE;
BEGIN
IF NOT c_high_sal%ISOPEN THEN
DBMS_OUTPUT.PUT_LINE('Cursor is Closed');
END IF;
OPEN c_high_sal;
IF c_high_sal%ISOPEN THEN
DBMS_OUTPUT.PUT_LINE('Cursor is open');
END IF;
LOOP
FETCH c_high_sal INTO high_sal;
IF c_high_sal%FOUND THEN
DBMS_OUTPUT.PUT_LINE(high_sal.employee_id||' '||high_sal.first_name
||' '||high_sal.last_name||' '||high_sal.salary);
ELSE
EXIT;
END IF;
END LOOP;
DBMS_OUTPUT.PUT_LINE(' Number of rows fetched : '||c_high_sal%ROWCOUNT);
CLOSE c_high_sal;
IF NOT c_high_sal%ISOPEN THEN
DBMS_OUTPUT.PUT_LINE('Cursor is closed ');
END IF;
END;
Cursor is Closed
Cursor is open
100 Steven King 24000
101 Neena Kochhar 17000
102 Lex De Haan 17000
145 John Russell 14000
146 Karen Partners 13500
201 Michael Hartstein 13000
205 Shelley Higgins 12000
108 Nancy Greenberg 12000
147 Alberto Errazuriz 12000
168 Lisa Ozer 11500
Number of rows fetched : 10
Cursor is closed
The cursor declared in the preceding example executes highest salary paid employees. Using sub-query in cursor. declare high_sal
with %ROWTYPE
.
The cursor is not yet open, %ISOPEN
attribute is used to see whether cursor is open with help of IF
-THEN
statement, where %ISOPEN
turn TRUE
.
Cursor is open, now to show with DBMS_OUTPUT
, %ISOPEN
is used with IF
-THEN
statement, where %ISOPEN
turn TRUE
.
Now loop start and cursor c_high_sal
is fetched into variable high_sal
, again %FOUND
attribute is used with IF
-THEN
-ELSE
statement to show the list of highest paid employees, where %FOUND
attribute turns true
till the last fetch of cursor and exits from the loop.
Once it is out of the loop, DBMS_OUTPUT
fetches the last number of row of loop and shows the figure with %ROWCOUNT
attribute. Cursor is closed.
After the cursor is closed, %ISOPEN
is used with the help of IF
-THEN
statement to show the cursor is closed.
Example 11. Cursor attributes with Implicit cursor
DECLARE
v_dept departments.department_id%TYPE := 270;
v_dept_name departments.department_name%TYPE;
BEGIN
DELETE FROM dept1 WHERE department_id = v_dept;
IF SQL%FOUND THEN
INSERT INTO dept1 VALUES(270,'Personnel',200,1700);
END IF;
DBMS_OUTPUT.PUT_LINE('Number of rows inserted : '||SQL%ROWCOUNT);
SELECT department_name INTO v_dept_name
FROM dept1
WHERE department_id = 270;
DBMS_OUTPUT.PUT_LINE('Department Name : '||v_dept_name);
END;
Number of rows inserted : 1
Department Name : Personnel
An implicit cursor can tell you how many rows were affected by an update
, insert
and delete
with the help of %FOUND
and %ROWCOUNT
attribute.
We have already created a table named dept1
in the previous example. In the following example, variables v_dept
and v_dept_name
are reference by table departments, as we are working on table dept1
,it will not give any error unless the structure or datatype is different from the other table, i.e., departments
and dept1
.
Deleting from dept1
table where department_id
is 270
.
Using SQL instead of cursor name as it is implicit cursor with no cursor name.
Use %FOUND
attribute with IF
-THEN
statement to insert row in dept1
table with new department_name
, i.e., Personnel
. Once it is inserted to know how many rows were affected, use %ROWCOUNT
in DBMS_OUTPUT
.
In same execution block using v_dept_name
, we are fetching new department name with select
statement.
I have not placed any commit because I don't want to update any row so later I will rollback it.
To learn more about implicit cursor and its attributes, refer to PL/SQL User's Guide and Reference chapter 6.
NOTE: Between implicit and explicit cursor, the fastest is implicit cursor, because opening and closing of cursor and looping is automatically handled by Oracle. Cursor for loop is also an implicit cursor where opening and closing of cursor in done implicitly. But cursor should be used according to requirement because each type of cursor has its own advantages and disadvantages. To know more about this, you can refer to Steven Feuerstein's book PL-SQL Programming 5th edition.
The concepts like Refs cursor, cursor with bulk operations, cursor with packages and cursor with exception are not covered which is out of scope of this article. I may write my next article covering all these concepts and thank you for reading this article.