Teach Yourself PL/SQL in 21 Days- P3

other-T ABLE 3.7 Expressions Using the AND Operator The Syntax for the OROperator boolean_expression OR boolean_expression In this syntax,boolean_expressioncan be any expression resultin

ANDTheANDoperator is used to join two comparison expressions when you are interested intesting whether both expressions are true It can also be used for the same purpose withtwo Boolean variables—to check to see whether both are equal to true

The Syntax for the ANDOperator

boolean_expression AND boolean_expression

In this syntax,boolean_expressioncan be any expression resulting in a Boolean, or

true/false, value This is often a comparison expression such as (a = b), but can also

be a variable of the BOOLEANdatatype

TheANDoperator returns a value of trueif both expressions each evaluate to true; wise, a value of falseis returned Use ANDwhen you need to test several conditions andexecute some code only when they are all true Table 3.7 shows some sample expres-sions using the ANDoperator

other-T ABLE 3.7 Expressions Using the AND Operator

The Syntax for the OROperator

boolean_expression OR boolean_expression

In this syntax,boolean_expressioncan be any expression resulting in a Boolean, or

true/false, value This is often a comparison expression such as (a = b), but can also

be a variable of the BOOLEANdatatype

TheORoperator returns a value of trueif any one of the expressions evaluates to true

A value of falseis returned only if both the expressions evaluate to false Table 3.8shows some sample expressions using the ORoperator

T ABLE 3.8 Expressions Using the OR Operator

(5 <> 5) OR (4 >= 100) OR (2 < 2) false (5 = 4) OR (5 = 5) true

‘Mon’ IN (‘Sun’,’Sat’) OR (2 = 2) true

String Operators

PL/SQL has two operators specifically designed to operate only on character string data.These are the LIKEoperator and the concatenation (||) operator The LIKEoperator is acomparison operator used for pattern matching and was described earlier in the sectiontitled “Comparison Operators,” so only the concatenation operator is described here

The Syntax for the Concatenation Operator

string_1 || string_2

In this syntax,string_1andstring_2are both character strings and can be either stringconstants, string variables, or a string expression The concatenation operator returns aresult consisting of all the characters in string_1followed by all the characters instring_2.

Listing 3.6 shows several ways in which you can use the concatenation operator

L ISTING 3.6 Use of the Concatenation Operator

1: Demonstrate the concatenation operator 2: DECLARE

3: a VARCHAR2(30);

4: b VARCHAR2(30);

5: c VARCHAR2(30);

6: BEGIN 7: Concatenate several string constants.

8: c := ‘Jack’ || ‘ AND ‘ || ‘Jill’;

Jack and Jill went up the hill,

to fetch a pail of water.

PL/SQL procedure successfully completed.

The preceding code shows the concatenation operator used in several differentways Notice that you do not always have to assign the result directly to a stringvariable For example, in line 13, the concatenation operator is used to create a stringexpression that is passed as input to the PUT_LINEprocedure

Using Comparison Operators with Strings

You can use any of the PL/SQL comparison operators to compare one character string toanother Strings can be compared for equality, for inequality, to see if one string is lessthan another, to see if one string matches a given pattern, and so on When using charac-ter strings in comparison expressions, the result depends on several things:

• Character set

• Datatype

• Case (upper versus lower)

The Effect of Character Set on String Comparisons

When comparing two strings to see if one is greater or less than another, the resultdepends on the sort order of the underlying character set being used In the typical ASCIIenvironment, all lowercase letters are greater than all uppercase letters, digits are lessthan all letters, and the other characters fall in various places depending on their corre-sponding ASCII codes However, if you were working in an EBCDIC environment, youwould find that all the digits were greater than the letters and all lowercase letters areless than all uppercase letters, so be careful

The Datatype’s Effect on String Comparisons

The underlying datatype has an effect when comparing two string variables, orwhen comparing a string variable with a constant Remember that variables oftheCHARdatatype are fixed-length and padded with spaces Variables of the VARCHAR2

datatype are variable-length and are not automatically padded with spaces When paring two CHARdatatypes, Oracle uses blank-padded comparison semantics This means

com-that Oracle conceptually adds enough trailing spaces to the shorter string to make itequal in length to the longer string and then does the comparison Trailing spaces alonewill not result in any differences being found between two springs Oracle also does thesame thing when comparing two string constants However, when one of the values in a

A NALYSIS

comparison is a variable-length string, Oracle uses non-padded comparison semantics The use of non-padded comparison semantics means that Oracle does not pad either of

the values with spaces, and any trailing spaces will affect the result Listing 3.7 showsseveral string comparisons that illustrate this point

L ISTING 3.7 Demonstration of String Comparison Semantics

1: Demonstrate string comparision semantics 2: DECLARE

10: IF ‘Jonathan’ = ‘Jonathan ‘ THEN 11: DBMS_OUTPUT.PUT_LINE

12: (‘Constant: ‘’Jonathan’’ = ‘’Jonathan ‘’’);

13: END IF;

14: Fixed length strings are also compared with blank-padded 15: comparison semantic, so the fact that one is longer doesn’t matter 16: fixed_length_10 := ‘Donna’;

17: fixed_length_20 := ‘Donna’;

18: IF fixed_length_20 = fixed_length_10 THEN 19: DBMS_OUTPUT.PUT_LINE(‘Char: ‘’’ || fixed_length_10 || ‘’’ = ‘’’ 20: || fixed_length_20 || ‘’’’);

Char and Varchar2: ‘Donna ‘ NOT = ‘Donna’

Both Varchar2: ‘Donna’ = ‘Donna’

PL/SQL procedure successfully completed.

You can see from the output that the first three comparisons in Listing 3.7 useblank-padded comparison semantics The strings being compared are considered

to be equal even though the number of trailing spaces differs in each case The fourthcomparison, however, compares a VARCHAR2variable against a CHARvariable Becauseone of the strings in question is variable-length, the trailing spaces count, and the twostrings are not considered to be equal

The Effect of Case on String Comparisons

PL/SQL string comparisons are always case-sensitive The obvious ramification of this isthat a lowercase string such as ‘aaa’is not considered equal to its uppercase equivalent

of‘AAA’ But case also makes a difference when comparing two strings to see which isgreater In an ASCII environment, the letter ‘A’will be less than the letter ‘B’ However,the letter ‘a’will not only be greater than ‘B’; it will be greater than ‘Z’

A NALYSIS

If you need to perform case-insensitive string comparisons, use PL/SQL’s built-in UPPER() function For example:

IF UPPER(‘a’) < UPPER(‘B’) THEN

You can use the LOWER() function in the same manner.

Tip

Use of Comparison Operators with Dates

Date comparison works pretty much as you might expect and has fewer complexitiesthan string comparisons do Earlier dates are considered to be less than later dates, and itfollows that more recent dates are greater than earlier dates The only complication arisesfrom the fact that PL/SQL date variables also contain a time component Listing 3.8illustrates this and another potential problem to be aware of when comparing date valuesagainst each other

L ISTING 3.8 Date Comparison Example

1: Demonstrate date comparisions 2: DECLARE

3: payment_due_date DATE;

4: BEGIN 5: In real life the payment_due date might be read from 6: a database or calculated based on information from a database.

7: payment_due_date := TRUNC(SYSDATE);

8: Display the current date and the payment date.

9: DBMS_OUTPUT.PUT_LINE(‘Today is ‘ || TO_CHAR(SYSDATE,’dd-Mon-yyyy’)); 10: DBMS_OUTPUT.PUT_LINE(‘Payment is due on ‘

11: || TO_CHAR(payment_due_date,’dd-Mon-yyyy’));

12: IF payment_due_date = SYSDATE THEN 13: DBMS_OUTPUT.PUT_LINE(‘Payment is due today.’);

14: ELSE 15: DBMS_OUTPUT.PUT_LINE(‘Payment can wait a while.’);

22: END IF;

23: END;

24: /

Today is 01-Jun-1997 Payment is due on 01-Jun-1997 Payment can wait a while.

Wrong! Payment is due today!

PL/SQL procedure successfully completed.

Today’s date and the payment due date both match, yet the IFstatement in line

12 failed to detect this Why? Because SYSDATEis a function that returns the rent date and time, with the time resolved down to the second The payment_due_datevariable will contain a time of midnight because the TRUNCfunction was used when the

cur-INPUT

A NALYSIS

due date was set See line 7 The only time that line 12 would function correctly would

be for one second at midnight each day In line 18, the TRUNCfunction is used to truncatethe time values from the two dates, resulting in a comparison that works as desired inthis case Admittedly this is a contrived example, but it does illustrate the problem

Be especially careful comparing dates read from the database Sometimes they have time components that you don’t expect Even if programmers aren’t supposed to store a time, it’s a pretty easy mistake to make.

Tip

Having the time as part of a date variable is not necessarily a bad thing It’s just thing you need to be aware of, especially when comparing dates with each other

some-Exploring Expressions

When you combine values and operators to produce a result, you have an expression

You have already learned about the various datatypes available in PL/SQL and PL/SQL’sextensive collection of operators In addition, in order to use expressions effectively inyour code, you also need to understand

• Operator precedence

• Use of parentheses

• Types of expressions

• The effects of null values in an expression

• Conversion between datatypesUnderstanding the effects of a null value on an expression is particularly important,especially when you move into retrieving data from a database The remainder of thischapter discusses each of these items in detail

Expressions Defined

Simply put, an expression is some combination of variables, operators, literals, and tions that returns a single value Operators are the glue that hold an expression togetherand are almost always present The other elements might not all be present in everyexpression

func-In its very simplest form, an expression might simply consist of a literal value, a variablename, or a function call The first few entries in Table 3.9 are examples of this type ofexpression More typical expressions involve two values and an operator, with the opera-

can be built up by stringing several simple expressions together with various operatorsand function calls Finally, the unary operators can be applied to any expression or value.

T ABLE 3.9 Sample Expressions

some_variable_name Evaluates to the contents of the variable

SYSDATE An Oracle function that returns the current date

1000 + 2000 A typical expression using a binary operator

-1000 An expression using a unary operator

10 * 20 + 30 / 2 Two expressions joined together

LENGTH(‘Lansing ‘ || ‘MI’) A function call evaluating a sub-expression is itself an

Operator Precedence

When evaluating an expression consisting of different values, datatypes, and operators,Oracle follows a specific set of rules that determine which operations are done first Eachoperator has an assigned precedence Operators with a higher precedence are evaluatedfirst Operators of the same precedence level are evaluated from left to right Table 3.10shows these precedence levels for each of the various operators

T ABLE 3.10 Operator Precedence

-used as unary operators)

Fourth + , - , || Addition, subtraction, and string

concate-nation Fifth = , <> , != , ~= , < , Comparison

> , <= , >= , LIKE ,

BETWEEN , IN ,

IS NULL

Take another look at the expression referred to in the previous section The following listshows the steps Oracle would take to evaluate it:

You can control the order in which Oracle evaluates an expression by using parentheses

Oracle will evaluate any part of an expression in parentheses first If parentheses arenested, Oracle will always evaluate the innermost expression first and then move out-wards Here is what happens to the preceding expression if you add some parentheses:

1 (1-5)**2<=10*(4-20)

2 (-4)**2<=10*(-16)

3 16<=-160

4 false

String expressions are those that return character strings as results, and date expressionsare those that result in a date-time value.

Generally speaking, you can use an expression of the appropriate datatype anywhere inyour PL/SQL code where a value is required The exception to this would be in functionand procedure calls that modify their arguments

Null Values in Expressions

Until now, the discussion has ignored the effect of nulls in expressions This was done inorder to concentrate on the normal function of each of the operators and also becausenulls pretty much have the same effect regardless of the operation being done

What is a null? The term is best understood as referring to an unknown value Any

vari-able or expression is considered null when the value of that varivari-able or expression is

unknown This situation can occur if you declare a variable and use it in an expressionwithout first assigning a value Because the variable has no assigned value, the result ofthe expression can’t be known More commonly, nulls are encountered when readingdata from a database Oracle, like any other relational database, does not force you to

Use parentheses in complex expressions, even when they are not strictly essary, in order to make the intended order of evaluation clear to other pro- grammers.

nec-Tip

store a value for each column in a table When no specific value is stored, the contents ofthat column are considered unknown, and the column is referred to as being null.The effects of nulls are particularly insidious when writing Boolean expressions,such as the WHEREclause in a SQL SELECTstatement SQL uses what is called

three-valued logic Three-valued logic says that the result of a Boolean expression can be

eithertrue,false, or NULL Many a programmer has felt the sting of an IFstatementgone awry because of an unexpected null value, and some consider three-valued logic to

be more of a three-pronged pitchfork prodding them in the behind The code in Listing3.9 shows why nulls can cause so much grief

L ISTING 3.9 Effects of Nulls on Boolean Expressions

1: Demonstrate the effects of null values on boolean expressions.

2: DECLARE 3: a INTEGER;

4: b BOOLEAN;

5: n INTEGER; this will be our null value.

6: BEGIN 7: Assign a value to the variable A, but leave N null.

8: a := 2;

9: Note that the test for A=N fails.

10: IF a = n THEN 11: DBMS_OUTPUT.PUT_LINE(‘a = n is true’);

12: ELSE 13: DBMS_OUTPUT.PUT_LINE(‘a = n is not true’);

14: END IF;

15: But also note that the test for a <> n fails

16: IF a <> n THEN 17: DBMS_OUTPUT.PUT_LINE(‘a <> n is true’);

18: ELSE 19: DBMS_OUTPUT.PUT_LINE(‘a <> n is not true’);

20: END IF;

21: Here is an expression that many people first 22: expect to always be true.

23: IF (a = n) OR (a <> n) THEN 24: DBMS_OUTPUT.PUT_LINE(‘(a = n) or (a <> n) is true’);

25: ELSE 26: DBMS_OUTPUT.PUT_LINE(‘(a = n) or (a <> n) is not true’);

27: END IF;

28: TRUE and NULL = NULL 29: IF (a = 2) AND (a <> n) THEN 30: DBMS_OUTPUT.PUT_LINE(‘TRUE and NULL = TRUE’);

31: ELSE 32: DBMS_OUTPUT.PUT_LINE(‘TRUE and NULL = NULL’);

33: END IF;

INPUT

continues

34: TRUE or NULL = TRUE 35: IF (a = 2) OR (a <> n) THEN 36: DBMS_OUTPUT.PUT_LINE(‘TRUE or NULL = TRUE’);

37: ELSE 38: DBMS_OUTPUT.PUT_LINE(‘TRUE or NULL = NULL’);

39: END IF;

40: NOT NULL = NULL 41: IF (NOT (a = n)) IS NULL THEN 42: DBMS_OUTPUT.PUT_LINE(‘NOT NULL = NULL’);

52: END IF;

53: TIP: a three-valued if construct.

54: b := (a <> n);

55: IF b THEN 56: DBMS_OUTPUT.PUT_LINE(‘a <> n is TRUE’);

57: ELSIF NOT b THEN 58: DBMS_OUTPUT.PUT_LINE(‘a <> n is FALSE’);

59: ELSE 60: DBMS_OUTPUT.PUT_LINE(‘a <> n is NULL’);

TRUE or NULL = TRUE NOT NULL = NULL The values are not equal.

a <> n is NULL PL/SQL procedure successfully completed.

Listing 3.9 is a somewhat contrived example, but it illustrates very well theeffects of nulls on comparison expressions Take a close look at what is going onhere The first IFstatement in line 10 tests for a = n As you might expect, this is nottrue, but it is important to understand that it is not falseeither The second IFstatement

in line 16 proves this The test there is for a <> n, the exact opposite of the previous

L ISTING 3.9 continued

A NALYSIS

• How nulls propagate in expressions

• How the logical operators AND,OR, and NOThandle nulls

• How the IFstatement deals with nulls

In an expression, null values propagate For the most part, any arithmetic, date, string, orBoolean expression containing even one nullvalue will also evaluate to null There aresome exceptions to this rule, which are described shortly

Where strings are concerned, Oracle considers an empty string (‘’ for ple) to be a null A string of spaces, however, is not the same as an empty string, and is not null.

exam-Note

The logical operators AND,OR, and NOTare often used to link together comparison sions Table 3.11 shows how these operators function in expressions with null values

expres-T ABLE 3.11 Three-Valued Logic Truth Table

Lines 28 through 43 in Listing 3.9 contain some IFstatements that demonstrate howeach of the logical operators operate on null values.

TheIFstatement is the fundamental decision-making structure of PL/SQL Give it aBoolean expression, and it evaluates that expression and makes a decision as to whichpiece of code to execute However, Boolean expressions can have three values:TRUE, FALSE, and NULL An IFstatement has only two parts: the code to be executed when anexpression is TRUE, and the code to be executed when it isn’t There is a mismatch here,and it’s very important to keep in mind that the ELSEportion will be executed when theresult of an expression is unknown, or in other words, when the expression is null

Lines 53 through 61 of Listing 3.9 show a way to construct an IF statement that has separate execution paths for TRUE , FALSE , and NULL

Tip

There are some exceptions to the general rule that nulls propagate in expressions Nullcharacter strings are sometimes handled as if they were zero-length strings, and PL/SQLdoes have some functions and operators that have been specifically designed to help youwork with nulls

You can concatenate strings, even if one is null, and get the results you would expect.This is because the concatenation operator simply ignores any null strings However, ifall the strings are null, the result will be null Also bear in mind that PL/SQL treats azero-lengthVARCHAR2string as a null value

Treating a zero-length string as a null value is an Oracle-specific behavior that is not specified in the ANSI standard for SQL.

Note

You can use the IS NULLoperator to see whether or not a particular variable or sion is null It allows your code to detect and act on null values You saw an example ofthis earlier in Listing 3.5 The IS NULLoperator returns only a trueorfalsevalue,never a NULL

Usually, when people use NVL, they simply supply a variable for the expression.

However, that doesn’t have to be the case When you invoke NVL, the result of the firstexpression is computed If that result is not null, then it will be returned as the result

ofNVL If the result of the expression does happen to be null, then NVLwillreturn the value of the second argument as its result The end result is that the

value_if_expression_is_nullargument becomes an alternate value for expression, to

be used if expressionis null

Lines 44 through 52 of Listing 3.9 show an interesting use of the NVLfunction to accountfor the possibility of the variable nbeing null You can read more about NVLin

Appendix B

The built-in SQL DECODEfunction actually treats null as a specific value instead of anunknown value It might seem contradictory, but it’s useful DECODEis also described inAppendix B

Always use the IS NULL operator when checking for null values Do not use the equality or inequality operators to compare a variable to null You can code a statement such as IF some_var = NULL, but you won’t get the results you might expect Use IF some_var IS NULL instead.

Do initialize all your variables in order to

eliminate the possibility of null values

retriev-ing values from that database in order to replace null values with an acceptable alternative.

Don’t forget to think through the

possi-ble implications of null values in every expression you write, especially the Boolean, including comparison and expressions.

Converting Datatypes

Sometimes you need to convert a value of one datatype to another This is frequently truewith dates and numbers, which are often converted to and from character strings Youmight want to display a date, for example, so you must first convert it to a characterstring in the desired format There are two ways of approaching the issue of conversion.One is to rely on Oracle to implicitly convert datatypes, which it will do automaticallywhen it makes sense The second and more preferred method is to code your conversionsexplicitly

Implicit Conversion

When you mix different datatypes in an expression, Oracle will convert them for

you when it makes sense to do so This is referred to as implicit conversion.

Listing 3.10 shows several examples of implicit conversion

L ISTING 3.10 Implicit Conversion Examples

1: Demonstrate implicit conversion 2: DECLARE

11: cd1 := ‘15-Nov-61’;

12: Now assign the string to a date variable.

13: The conversion is implicit.

14: d1 := cd1;

15: Now assign that date variable to another string.

16: Again the conversion 17: is implicit, but this time the conversion is 18: from a date to a string.

CD1 = 15-Nov-61 CD2 = 15-NOV-61 CN1 = 995 CN2 = 995.99 PL/SQL procedure successfully completed

The code in Listing 3.10 illustrates some common implicit conversions The firstassignment, in line 11, causes no conversion at all because a string is assigned to

a string variable The assignment statement in line 14, however, does represent an

implic-it conversion because implic-it must convert the string representation of the date to Oracle’sinternal format before it can assign the value to d1 In line 19, that date is again convert-

ed back to a string format Lines 23 through 28 repeat the same process, but this timewith a number

Implicit conversions are convenient, but can sometimes cause problems In relying onthem, you are relying on Oracle’s built-in assumptions and on default settings you mightnot even be aware of, and which might change from one release to another The format

of a date leads to some good examples Did you know that Oracle’s default date formatvaries depending on the language setting? That it can also be installation-dependent? Andthat it can vary between a client PC executing a Developer/2000 script and a databaseserver? In fact, the date format can even be changed for the duration of a particular ses-sion Figure 3.2 illustrates this by showing the same PL/SQL code succeeding once andthen failing after the date format has been changed

for-For the reasons just listed, it is often safer to code conversions explicitly Explicit sions also better document your code by making it clear to other programmers exactlywhat is happening

TO_DATE Converts a character string to a date

TO_NUMBER Converts a character string to a number

TO_CHAR Converts either a number or a date to a character string

Each of these functions takes three arguments: the value to be converted, a format stringspecifying how that conversion is to take place, and optionally a string containing language-specific parameters These functions are described in detail on Day 6, butListing 3.11 gives some common examples of how you can use them

L ISTING 3.11 Examples of the Conversion Functions

1: Demonstrate conversion functions 2: DECLARE

12: d1 := TO_DATE(‘1/1/02’,’mm/dd/yy’);

13: d2 := TO_DATE(‘1-1-1998’,’mm-dd-yyyy’);

14: d3 := TO_DATE(‘Jan 1, 2000’,’mon dd, yyyy’);

15: Year 2000 problems? Note the effect of using rr instead of yy.

n1 = 123.99 n2 = $1,235.95 PL/SQL procedure successfully completed.

Lines 12 through 16 show the TO_DATEfunction being used to convert somecommon date formats to a date variable Lines 17 through 20 display these datesand show some more formatting possibilities Lines 22 through 25 show some examples

of conversions between numeric and character datatypes

Summary

Today you have learned about writing PL/SQL expressions You have read descriptions

of each of the PL/SQL operators and seen examples of these operators in action Youhave also seen how to write complex expressions and how the rules of operator prece-dence govern Oracle’s evaluation of these expressions Remember that you can useparentheses when you need to exercise control over a calculation Most important toremember are the effects of null, or unknown, values on expressions This is a particular-

ly important subject to keep in mind when writing comparisons for use with ifments Mastering this one area will save you untold grief as you write code in the future

state-Q&A

Q Why does the expression TRUE AND NULLevaluate to NULL, but the expression

TRUE OR NULLevaluates to true?

A This is a good question To understand the answer, it might help to think in terms

of null being an unknown value The ANDoperator requires that both its operands

be true in order to return a trueresult If one of the operands is unknown, youcan’t be sure that if it were known it would be true, so ANDmust evaluate to false

in this case Things are different, however, for the expression TRUE OR NULL The

ORoperator only requires one of its operands to be truein order to return a true

result In the case of TRUE OR NULL, you do know that one operand is true

A NALYSIS

Whether the other operand is true,false, or unknown doesn’t matter at this pointbecause you have one you know is true, and one is all you need.

Q Does the INoperator let me do anything that I couldn’t do otherwise?

A No, not really, but it does make your code more readable and easier to maintain.

The expression x IN (1,3,4,10,30,30,40,100)is equivalent to x=1 OR x=3 ORx=4 OR x=10 OR x=30 OR x=40 OR x=100, but you will probably find the firstversion a bit easier to read and understand

Q You said that a statement like IF X = NULL THEN would not work as expected, and that IF X IS NULL THEN should be used instead Why?

A The first expression will never be true It will always evaluate to nullbecause one

of the operands is null, and it can never be known if two values are equal when one

of the values is unknown The second expression uses the IS NULLoperator, which

is designed to check for nulls It specifically checks to see if the value of X isunknown and evaluates to trueif that is the case

Q When I am comparing strings, especially when comparing a CHARstring to a

VARCHAR2string, is there a convenient way to tell PL/SQL to ignore any ing spaces in the CHARstring?

trail-A Yes Use the built-in RTRIMfunction For example:IF RTRIM(char_string) = varchar2_string then

Q I’m comparing two dates and only want to know if they are in the same year Can I use the TRUNCfunction to accomplish this?

A Yes By default, the TRUNCfunction truncates the time portion of a date, but theoptional second argument enables you to specify a different point of truncation Tocompare only the years, you can write:IF TRUNC(date_1,’yyyy’) =

TRUNC(date_2,’yyyy’) THEN

Q Sometimes you capitalize your date format strings Why?

A When converting a date to a character string for display purposes, capitalizing parts

of the format string controls whether that part of the date is capitalized Supposethe current month is January The expression TO_CHAR(SYSDATE,’mon’)wouldresult in a value of ‘jan’, the expression TO_CHAR(SYSDATE,’Mon’)would result in

a value of ‘Jan’, and the expression TO_CHAR(SYSDATE,’MON’)would result in avalue of ‘JAN’

Q Is there any way to prevent Oracle from implicitly converting data from one type to another?

A No, there is not The only way that you can prevent implicit conversions from

hap-pening is to watch what you’re doing and code all conversions explicitly

1 What is the difference between a unary operator and a binary operator?

2 What are the results of each of the following expressions?

• Functions

• TheNOCOPYhint

• TheNULLstatement

• TheIFstatement

• NestedIFs

• TheELSIFstatement

• FORloops

Exploring PL/SQL Functions

As you saw in Day 1, “Learning the Basics of PL/SQL,” functions are similar to PL/SQLprocedures except for the following differences:

• Functions return a value

• Functions are used as part of an expression

Why should you write functions? You have many reasons The main reason is to reducethe total lines of coding and take a modular approach to writing code You can retype ineach PL/SQL block the same repetitive lines of code, or you can write a function Then,you can call that function repeatedly as needed What if you had to change all thoseblocks of code for one small reason? Just trying to find and change all those blockswould make a COBOL programmer shudder when contemplating Year 2000 changes!With functions, you simply make the change in one location (Keep in mind that if theparameters to be passed to the function change, you still have some editing to do withinthe PL/SQL blocks.)

Even if you do not write your own functions, don’t forget that Oracle provides you a vastarray of powerful built-in functions However, if you do not get the opportunity to writeyour own functions, you miss out on a powerful feature of PL/SQL

In Day 1, you created a function called SS_THRESH, which is shown in Listing 4.1 Thisfunction simply returns a value formatted as a number with nine total digits, two ofwhich are allocated to the decimal place Your values range from –9999999.99to9999999.99.

L ISTING 4.1 The SS_THRESH Function

CREATE OR REPLACE FUNCTION ss_thresh RETURN NUMBER AS

The Syntax for Declaring a Function

A function is declared as follows:

FUNCTION function_name [(parameters)]

RETURN return_data_type IS|AS

This syntax has the following statements and parameters:

• function_name—The function name follows the keyword FUNCTIONand followsthe standard naming convention, which is covered in Day 2, “Writing Declarationsand Blocks.”

• parameters—Functions let you pass parameters in and out The brackets []cate it is optional to include parameters

indi-• RETURN—This is the type of data returned You can have more than one RETURN

statement in a function, but only one is executed One RETURNstatement is requiredeven if you return nothing or ignore what is returned

• IS|AS—These parameters allow you to set up variables local to the function Thesevariables can hold values passed from the parameters or values assigned in thefunction The variables can’t be seen outside of the function

• BEGIN—This statement starts the execution of statements pertinent to the function

• EXCEPTION—Again, this statement is optional, but it allows you to handle thePL/SQL block properly when an error occurs If you do not address this up front,when an error occurs, control is passed back to the PL/SQL block with the originalvalues instead of the values that should have been returned

• END—This statement denotes the end of the function Errors will occur if you omittheENDstatement The function name may follow the ENDstatement If the name isincluded, it must match the name exactly as the name of the function