SQL Intermediate
Aggregate Functions
• •
2
3
Q1. What will be displayed by the following SQL statement?
SELECT count(*), count(mark) FROM enrolment;
8, 8 8, 3 3, 3 3, 8
4
Q2. What will be displayed by the following SQL statement?
SELECT count(*), count(stu_nbr), count(distinct stu_nbr) FROM enrolment;
8, 8, 4 8, 8, 8 8, 4, 8 8, 4, 4
5
Q3. We want to calculate the average mark of the 8 rows in the above table. What SQL statement should we use?
Note: We want to calculate (78+35+65)/8=22.25
SELECT avg(mark) FROM enrolment;
SELECT sum(mark)/count(mark) FROM enrolment; SELECT sum(mark)/count(*) FROM enrolment; SELECT avg(NVL(mark,0)) FROM enrolment; None of the above.
More than one option is correct.
Anatomy of an SQL Statement – Revisited
clauses
statement
Predicate / search condition
6
GROUP BY
▪ If a GROUP BY clause is used with aggregate function, the DBMS will apply the aggregate function to the different groups defined in the clause rather than all rows.
SELECT avg(mark) FROM enrolment;
SELECT unit_code, avg(mark) FROM enrolment
GROUP BY unit_code ORDER BY unit_code;
7
8
What output is produced?
SELECT avg(mark) FROM enrolmentA;
SELECT unit_code, avg(mark) FROM enrolmentA
GROUP BY unit_code ORDER BY unit_code;
SELECT unit_code, avg(mark), count(*) FROM enrolmentA
GROUP BY unit_code
ORDER BY unit_code;
Unit_code
Mark
Studid
Year
FIT2094
80
111
2016
FIT2094
20
111
2015
FIT2004
100
111
2016
FIT2004
40
222
2015
FIT2004
40
333
2015
9
Unit_code Mark
FIT2094 80 FIT2094 20 FIT2004 100 FIT2004 40 FIT2004 40
Studid Year
111 2016 111 2015 111 2016 222 2015 333 2015
10
What output is produced?
Unit_code
Mark
Studid
Year
FIT2094
80
111
2016
FIT2094
20
111
2015
FIT2004
100
111
2016
FIT2004
40
222
2015
FIT2004
40
333
2015
SELECT unit_code, avg(mark), count(*) FROM enrolmentA
GROUP BY unit_code, year
ORDER BY unit_code, year;
11
Note: attributes in the GROUP BY clause do not have to appear in the select list
Unit_code Mark
FIT2094 80 FIT2094 20 FIT2004 100 FIT2004 40 FIT2004 40
Studid Year
111 2016 111 2015 111 2016 222 2015 333 2015
12
HAVING clause
▪ It is used to put a condition or conditions on the groups defined by GROUP BY clause.
SELECT unit_code, count(*) FROM enrolment
GROUP BY unit_code HAVING count(*) > 2;
13
What output is produced?
SELECT unit_code, avg(mark), count(*) FROM enrolmentA
GROUP BY unit_code
HAVING count(*) > 2
ORDER BY unit_code;
SELECT unit_code, avg(mark), count(*) FROM enrolmentA
GROUP BY unit_code
HAVING avg(mark) > 55
ORDER BY unit_code;
Unit_code
Mark
Studid
Year
FIT2094
80
111
2016
FIT2094
20
111
2015
FIT2004
100
111
2016
FIT2004
40
222
2015
FIT2004
40
333
2015
14
Unit_code
Mark
Studid
Year
FIT2094
80
111
2016
FIT2094
20
111
2015
FIT2004
100
111
2016
FIT2004
40
222
2015
FIT2004
40
333
2015
15
HAVING and WHERE clauses
SELECT unit_code, count(*) FROM enrolment
WHERE mark IS NULL GROUP BY unit_code HAVING count(*) > 1;
• The WHERE clause is applied to ALL rows in the table.
• The HAVING clause is applied to the groups defined by the GROUP BY clause.
• The order of operations performed is FROM, WHERE, GROUP BY, HAVING and then ORDER BY.
• On the above example, the logic of the process will be:
• All rows where mark is NULL are retrieved. (due to the WHERE clause)
• The retrieved rows then are grouped into different unit_code.
• If the number of rows in a group is greater than 1, the unit_code and the total is displayed. (due to the HAVING clause)
16
What output is produced?
Unit_code
Mark
Studid
Year
FIT2094
80
111
2016
FIT2094
20
111
2015
FIT2004
100
111
2016
FIT2004
40
222
2015
FIT2004
40
333
2015
SELECT unit_code, avg(mark), count(*) FROM enrolmentA
WHERE year = 2015
GROUP BY unit_code
HAVING avg(mark) > 30 ORDER BY avg(mark) DESC;
17
Unit_code
Mark
Studid
Year
FIT2094
80
111
2016
FIT2094
20
111
2015
FIT2004
100
111
2016
FIT2004
40
222
2015
FIT2004
40
333
2015
18
19
Unit_code
Mark
Studid
Year
FIT2094
80
111
2016
FIT2094
20
111
2015
FIT2004
100
111
2016
FIT2004
40
222
2015
FIT2004
40
333
2015
Q4. What is the output for:
SELECT unit_code, studid, avg(mark)
FROM enrolmentA
GROUP BY unit_code HAVING avg(mark) > 55 ORDER BY unit_code, studid;
A. FIT2094, 111, 50
B. FIT2004, 111, 60
C. FIT2004, 111, 60, 222, 333
D. FIT2004, 111, 100
E. Will print three rows
F. Error
Unit_code Mark
FIT2094 80 FIT2094 20 FIT2004 100 FIT2004 40 FIT2004 40
Studid Year
111 2016 111 2015 111 2016 222 2015 333 2015
20
SELECT stu_lname, stu_fname, avg(mark) FROM enrolment e JOIN student s
ON s.stu_nbr = e.stu_nbr GROUP BY s.stu_nbr;
The above SQL generates error message
Why and how to fix this?
• Why? Because the grouping is based on the stu_nbr, whereas the display is based on stu_lname and stu_fname. The two groups may not have the same members.
• How to fix this?
• Include the stu_lname,stu_fname as part of the GROUP BY condition.
• Attributes that are used in the SELECT, HAVING and ORDER BY must be included in the GROUP BY clause.
21
Subqueries
•
22
Types of Subqueries
Single-value
returns
Multiple-row subquery (a list of values – many rows, one column)
returns APPLE PEAR
Multiple-column subquery (many rows, many columns)
Main query
Subquery
APPLE
Main query
Subquery
Main query
returns
APPLE 4.99 PEAR 3.99
Subquery
23
24
Q5. What will be returned by the inner query?
SELECT *
FROM enrolment
WHERE mark > (SELECT avg(mark)
FROM enrolment GROUP BY unit_code);
A value (a single column, single row). A list of values.
Multiple columns, multiple rows. None of the above.
25
26
Q6. What will be returned by the inner query?
SELECT unit_code, stu_lname, stu_fname, mark FROM enrolment e join student s
on e.stu_nbr = s.stu_nbr
WHERE (unit_code, mark) IN (SELECT unit_code, max(mark)
FROM enrolment GROUP BY unit_code);
A value (a single column, single row). A list of values.
Multiple columns, multiple rows. None of the above.
Comparison Operators for Subquery
• Operator for single value comparison. =, <, >
• Operator for multiple rows or a list comparison. –equality
• IN –inequality
•ALL, ANY combined with <, >
27
28
Q7. Which row(s) in ENROL2 table will be retrieved by the following SQL statement?
SELECT * FROM enrol2
WHERE mark IN (SELECT max(mark)
FROM enrol2
GROUP BY unit_code);
A. 1, 2, 7 B. 7
C. 2, 3, 7
29
30
Q8. Which row/s in ENROL2 will be retrieved by the following SQL statement?
SELECT * FROM enrol2
WHERE mark > ANY (SELECT avg(mark)
FROM enrol2
GROUP BY unit_code);
1, 2, 3, 6, 7
2, 3, 7
3, 7
No rows will be returned
31
32
Q9. Which row/s in ENROL2 will be retrieved by the following SQL statement?
SELECT * FROM enrol2
WHERE mark > ALL (SELECT avg(mark)
FROM enrol2
GROUP BY unit_code);
1, 2, 3, 6, 7 2, 3, 7
3, 7
33
34
Q10. Find all students whose mark in any enrolled unit is lower than Wendy Wheat’s lowest mark for all units she is enrolled in. What would be a possible inner query statement for the above query (assume Wendy Wheat’s name is unique)?
A. SELECT min(mark) FROM enrol2
WHERE stu_lname=’Wheat’ AND stu_fname=’Wendy’;
B. SELECT min(mark)
FROM enrol2 e JOIN student s on e.studid = s.studid WHERE stu_lname=’Wheat’ AND stu_fname=’Wendy’;
C. SELECT min(mark) FROM enrol2;
D. SELECT mark
FROM enrol2 e JOIN student s on e.studid = s.studid WHERE stu_lname=’Wheat’ AND stu_fname=’Wendy’;
Summary
▪ Aggregate Functions
–count, min, max, avg, sum
▪ GROUP BY and HAVING clauses. ▪ Subquery
–Inner vs outer query
–comparison operators (IN, ANY, ALL)
35
PART 2
PL/SQL – Triggers (FIT3171)
36
Oracle Triggers
▪ A trigger is PL/SQL code associated with a table, which performs an action when a row in a table is inserted, updated, or deleted.
▪ Triggers are used to implement some types of data integrity constraints that cannot be enforced at the DBMS design and implementation levels
▪ A trigger is a stored procedure/code block associated with a table
▪ Triggers specify a condition and an action to be taken whenever that
condition occurs
▪ The DBMS automatically executes the trigger when the condition is met (“fires”)
▪ A Trigger can be ENABLE’d or DISABLE’d via the ALTER command
– ALTER TRIGGER trigger_name ENABLE;
37
Oracle Triggers – general form
CREATE [OR REPLACE] TRIGGER
{UPDATE | INSERT | DELETE}
[OF
[WHEN]
DECLARE BEGIN END;
…. trigger body goes here …..
38
Triggering Statement
BEFORE|AFTER INSERT|UPDATE [of colname]|DELETE ON Table ▪ The triggering statement specifies:
– –
the type of SQL statement that fires the trigger body.
the possible options include DELETE, INSERT, and UPDATE. One, two, or all three of these options can be included in the triggering statement specification.
the table associated with the trigger.
– if a triggering statement specifies UPDATE, an optional list of columns
–
▪ Column List for UPDATE
can be included in the triggering statement.
– if you include a column list, the trigger is fired on an UPDATE statement
only when one of the specified columns is updated.
– if you omit a column list, the trigger is fired when any column of the associated table is updated
39
Trigger Body
BEGIN …..
END;
▪ is a PL/SQL block that can include SQL and PL/SQL statements. These statements are executed if the triggering statement is issued and the trigger restriction (if included) evaluates to TRUE.
▪ Within a trigger body of a row trigger, the PL/SQL code and SQL statements have access to the old and new column values of the current row affected by the triggering statement.
▪ Two correlation names exist for every column of the table being modified: one for the old column value and one for the new column value.
40
Correlation Names
▪ Oracle uses two correlation names in conjunction with every column value of the current row being affected by the triggering statement. These are denoted by:
OLD.ColumnName & NEW.ColumnName
• For DELETE, only OLD.ColumnName is meaningful
• For INSERT, only NEW.ColumnName is meaningful
• For UPDATE, both are meaningful
▪ A colon must precede the OLD and NEW qualifiers when they are used in a trigger’s body, but a colon is not allowed when using the qualifiers in the WHEN clause.
▪ Old and new values are available in both BEFORE and AFTER row triggers.
41
FOR EACH ROW Option
▪ The FOR EACH ROW option determines whether the trigger is a row trigger or a statement trigger. If you specify FOR EACH ROW, the trigger fires once for each row of the table that is affected by the triggering statement. The absence of the FOR EACH ROW option means that the trigger fires only once for each applicable statement, but not separately for each row affected by the statement.
CREATE OR REPLACE TRIGGER display_salary_increase AFTER UPDATE OF empmsal ON employee
FOR EACH ROW
WHEN (new.empmsal > 1000)
BEGIN
DBMS_OUTPUT.PUT_LINE (‘Employee: ‘|| :new.empno ||’ Old salary: ‘|| :old.empmsal || ‘ New salary: ‘|| :new.empmsal);
END;
42
Statement Level Trigger
▪ Executed once for the whole table but will have to check all rows in the table.
▪ In many cases, it will be inefficient.
▪ No access to the correlation values :new and :old.
43
Oracle Data FK Integrity
▪ Oracle offers the options: – UPDATE
• no action (the default – not specified) – DELETE
• no action (the default – not specified) • cascade
• set null
▪ Subtle difference between “no action” and “restrict”
– RESTRICT – will not allow action if child records exist, checks first
– NO ACTION – allows action and any associated triggers, then checks integrity
▪ Databases implementations vary, for example:
– Oracle no RESTRICT
– IBM DB2, SQLite implement both as above
44
Common use of triggers
▪ In the model above OWNER is the PARENT (PK end) and VEHICLE is the CHILD (FK end)
▪ What should the database do to maintain integrity if the user:
– attempts to UPDATE the owner_no of the owner (parent)
– attempts to DELETE an owner who still has vehicles in the vehicle table
▪ Oracle, by default, takes the safe approach
– UPDATE NO ACTION (no update of PK permitted if child records)
– DELETE NO ACTION (no delete permitted if child records)
– what if you as the developer want UPDATE CASCADE?
45
Oracle Triggers
CREATE OR REPLACE TRIGGER Owner_Upd_Cas
BEFORE UPDATE OF owner_no ON owner
FOR EACH ROW
BEGIN
UPDATE vehicle
SET owner_no = :new.owner_no
WHERE owner_no = :old.owner_no;
DBMS_OUTPUT.PUT_LINE (‘Corresponding owner number in the VEHICLE
table has also been updated’);
END; /
▪ SQL Window: To CREATE triggers, include the RUN command (/) after the last line of the file
Implement UPDATE CASCADE rule
OWNER 1 —- has — M VEHICLE
:new.owner_no – value of owner_no after update :old.owner_no – value of owner_no before update
46
Common use of triggers – data integrity
▪ A trigger can be used to enforce user-defined integrity by triggering on a preset condition, carrying out some kind of test and then if the test fails, the trigger can raise an error (and stop the action) via a call to raise_application_error
The syntax for this call is:
the -20000 is the error number which is reported to the user, the error message is the error message the user will see. The error number can be any number less than or equal to -20000.
47
Common use of triggers – data integrity – example
For example: a trigger which will ensure any unit added (ie. inserted) to the UNIT table has a unit code which starts with ‘FIT’. Test your trigger and ensure it works correctly and shows your error message.
CREATE OR REPLACE TRIGGER check_unit_code BEFORE
INSERT ON unit
FOR EACH ROW
BEGIN
IF :new.unit_code NOT LIKE ‘FIT%’ THEN
raise_application_error(-20000, ‘Unit code must begin with FIT’);
END IF; END;
/
— Test Harness
— display before value select * from unit;
insert into unit values (‘ABC0001′,’Test Insert’,6);
— display after value select * from unit;
— closes transaction rollback;
48
Mutating Table
▪ A table that is currently being modified through an INSERT, DELETE or UPDATE statement SHOULD NOT be read from or written to because it is in a transition state between two stable states (before and after) where data integrity can be guaranteed.
– Such a table is called mutating table.
49
Triggers Case Study
50
• The student enrolment database contains two derived attributes no_student (total number of students) and ave_mark (average mark) .
• The total number of students is updated when an enrolment is added or deleted.
• The average mark is updated when an update on attribute mark is performed.
• For audit purpose, any deletion of enrolment needs to be recorded in an audit table. The recorded information includes the username who performed the deletion, the date and time of the deletion, the student no and unit code.
51
Q5. Based on the rule to maintain the integrity of the no_student attribute in the UNIT table as well as keeping the audit record, a trigger needs to be created for ________________ table. The trigger will update a value on ______________ table and insert a row to ________________ table.
A. UNIT, ENROLMENT, AUDIT
B. ENROLMENT, UNIT, AUDIT
C. STUDENT, ENROLMENT, AUDIT
D. AUDIT, UNIT, ENROLMENT
52
CREATE OR REPLACE TRIGGER triggername
BEFORE|AFTER INSERT|UPDATE [of colname]|DELETE [OR
…] ON Table
FOR EACH ROW
DECLARE
var_name datatype [, …]
BEGIN
….. END;
53
Q6. What would be an appropriate condition for the trigger described on the previous slide?
A. BEFORE INSERT OR DELETE ON enrolment.
B. AFTER INSERT OR DELETE ON enrolment.
C. BEFORE UPDATE OF mark ON enrolment.
D. AFTER UPDATE OF mark ON enrolment.
54
CREATE OR REPLACE TRIGGER change_enrolment
AFTER INSERT OR DELETE ON ENROLMENT
FOR EACH ROW
DECLARE
??????
BEGIN
END;
????????
55
Q7. What would be the logic to update the no_student attribute in the UNIT table when a new row is inserted to ENROLMENT?
A. UPDATE unit
SET no_student = no_student + 1
WHERE unit_code = unit code of the inserted row
B. UPDATE unit
SET no_student = (SELECT count (stu_nbr)
FROM enrolment
WHERE unit_code= unit code of the inserted row)
WHERE unit_code = unit code of the inserted row
C. UPDATE unit
SET no_student = no_student -1
WHERE unit_code = unit code of the inserted row
D. UPDATE unit
56
CREATE OR REPLACE TRIGGER change_enrolment
AFTER INSERT OR DELETE ON ENROLMENT
FOR EACH ROW
DECLARE
??????
BEGIN
IF INSERTING THEN
UPDATE unit
SET no_student = no_student + 1
WHERE unit_code = :new.unit_code
ENDIF;
?????
END;
57
Q8. What would be the logic for the trigger to deal with a deletion of a row in enrolment? Assume that a table audit_trail contains audit_time, user, sno and unitcode attributes.
A. UPDATE unit
SET no_student = no_student -1 WHERE unit_code = :old.unit_code;
B. INSERT INTO audit_trail VALUES (SYSDATE, USER,
:old.stu_nbr, :old.unit_code);
C. UPDATE unit
SET no_student = no_student – 1 WHERE unit_code = :new.unit_code;
D. a and b.
E. b and c.
58
CREATE OR REPLACE TRIGGER change_enrolment AFTER INSERT OR DELETE ON ENROLMENT
FOR EACH ROW
BEGIN
IF INSERTING THEN
UPDATE unit
SET no_student = no_student + 1 WHERE unit_code = :new.unit_code;
END IF;
IF DELETING THEN
UPDATE unit
SET no_student = no_student -1 WHERE unit_code = :old.unit_code;
INSERT INTO audit_trail VALUES (SYSDATE, USER, :old.stu_nbr, :old.unit_code);
END IF; END;
59
Test Harness
▪ it is not sufficient to code a trigger only, a suitable test harness must be developed at the same time and used to ensure the trigger is working correctly.
— display before value
select * from unit;
— test the trigger for insertion
insert into enrolment values (11111111,’FIT2001’,2013,2,null,null);
— display after value
select * from unit;
— test the trigger for deletion
delete from enrolment where stu_nbr = 11111111 and unit_code = ’FIT2001’and enrol_year =
2013 and enrol_semester = 2;
— display after value
select * from unit; select * from audit_trail; — closes transaction
rollback;
60
Statement Level Trigger
create or replace
TRIGGER DELETE_STATEMENT
AFTER DELETE ON ENROLMENT
BEGIN
INSERT INTO enrol_history VALUES (SYSDATE, USER, ‘Deleted’);
END;
Row Level Trigger
create or replace
TRIGGER DELETE_ENROLMENT
AFTER DELETE ON ENROLMENT
FOR EACH ROW
BEGIN
INSERT INTO audit_trail VALUES
(SYSDATE, USER, :old.stu_nbr, :old.unit_code);
END;
61
Oracle Triggers
▪ Use triggers where:
– a specific operation is performed, to ensure related actions are also performed
– to enforce integrity where data has been denormalised
– to maintain an audit trail
– global operations should be performed, regardless of who performs the operation
– they do NOT duplicate the functionality built into the DBMS
– their size is reasonably small (< 50 - 60 lines of code) ▪ Do not create triggers where:
– they are recursive
– they modify or retrieve information from triggering tables
62