SQL$HELP_OLD72.HLB  —  Conditional Expressions
    A conditional expression is an advanced form of the value
    expression that allows applications to return alternate
    information within an expression. See Value_Expressions for more
    information on the syntax. The following table describes the
    conditional expressions that are supported by Oracle Rdb.

    Table 3 Conditional Expressions

    Expression
    Name       Description

    ABS        ABS returns the absolute value of n.
    CASE       CASE alters the result of an expression. CASE can also
               generate or convert null values.
    COALESCE   COALESCE returns the first non-NULL value from a series
               of value expressions; otherwise, returns NULL.
    DECODE     Alternate form of the Case expression.
    GREATEST   GREATEST returns the greatest non-null value.
    LEAST      LEAST returns the least non-null value.
    NULLIF     NULLIF substitutes NULL when two value expressions are
               equal; otherwise, returns the first value.
    NVL        NVL returns the first non-NULL value from a series of
               value expressions; otherwise, returns NULL. NVL is a
               synonym for COALESCE.
    NVL2       If the first value expression to NVL2 is not NULL, then
               return the second value expression; otherwise return
               the third value expression.
    SIGN       SIGN returns the sign of the value

1  –  ABS Function

    The ABS function returns NULL if the passed value expression
    evaluates to NULL. The datatype of the result is the same as the
    passed value expression and supports scaled values of these data
    types: TINYINT, SMALLINT, INTEGER, BIGINT, REAL, FLOAT, DOUBLE
    PRECISION, INTERVAL, DECIMAL, NUMERIC and NUMBER.

    The absolute value function (ABS) returns NULL if the value
    expression evaluates to NULL. If the value expression evaluates
    to a value less than zero then that value is negated so that
    a positive value is returned. Otherwise the value is returned
    unchanged. For instance, ABS (-1) will return the value 1.

    ABS (a) is equivalent to the CASE expression

        case
            when a < 0 then - a
            else a
        end

    Example: Using the ABS function on an INTERVAL result of a date
    subtraction.

    SQL> select
    cont>    ABS ((birthday - current_date) year(3))
    cont> from employees
    cont> order by employee_id
    cont> limit to 10 rows;

      054
      047
      047
      064
      068
      062
      044
      069
      050
      074
    10 rows selected

    Example: Using ABS within a statistical function

    SQL> -- what is the average time in a job for each employee
    SQL> -- exclude anyone on there first job
    SQL> select
    cont>    employee_id,
    cont>    AVG (ABS (EXTRACT (MONTH FROM (job_start - job_end) month (4))))
    cont>           as "Average Job" edit using '--,---,--9.99" years"'
    cont> from JOB_HISTORY
    cont> where employee_id < '00200'
    cont> group by employee_id
    cont> having COUNT (*) > 1;
     EMPLOYEE_ID           Average Job
     00164                 14.00 years
     00165                 22.67 years
     00166                 20.00 years
     00167                 14.50 years
     00168                 26.33 years
     00169                 22.67 years
    ...etc...
     00197                 26.33 years
     00198                 37.00 years
     00199                 35.00 years
    30 rows selected
    %RDB-I-ELIM_NULL, null value eliminated in set function

2  –  COALESCE Expressions

    The COALESCE and NVL expressions return the first non-NULL value
    from a series of value expressions.

    SQL evaluates each value expression in a COALESCE or NVL
    expression until it can return a non-NULL value. If all the
    columns specified in the COALESCE or NVL expression contain NULL
    values, then NULL is returned.

    The data type of the resulting expression is a common data type
    to which all value expressions in the list can be converted.
    For example, COALESCE(SALARY_AMOUNT, ESTIMATED_BONUS, 1.23E+5)
    results in a DOUBLE PRECISION result because at least one
    argument is a floating point value.

    The following example replaces the stored NULL value in the
    MIDDLE_INITIAL column of the EMPLOYEES table with a hyphen:

    SQL> SELECT FIRST_NAME, LAST_NAME, MIDDLE_INITIAL,
    cont> COALESCE(MIDDLE_INITIAL, '-')
    cont> FROM EMPLOYEES
    cont> WHERE LAST_NAME LIKE 'L%';
     FIRST_NAME   LAST_NAME        MIDDLE_INITIAL
     Jo Ann       Lapointe         C                C
     Hope         Lapointe         NULL             -
     Stan         Lasch            P                P
     Norman       Lasch            NULL             -
     Peter        Lengyel          A                A
     Peter        Lonergan         V                V
    6 rows selected

3  –  CASE Expressions

    There are many situations where you might find it useful to alter
    the result of an expression. For example, you might have a table
    column called WORK_STATUS containing the data 0, 1, and 2 meaning
    Inactive, Full time, and Part time, respectively. The single
    character is more efficient to store than the definition of the
    character in the database. However, the definition of the single
    character is not always intuitive.

    There may also be times when you want to generate null
    values based on the information derived from the database or,
    conversely, convert a null value into a more concrete value like
    zero (0).  The CASE expressions provide an easy solution to these
    problems.

    There are two types of CASE expressions:

    o  Simple-matches two value expressions for equality

    o  Searched-allows complex predicate, including subqueries

    An example of the simple case expression follows:

    SQL> SELECT LAST_NAME, FIRST_NAME,
    cont>   CASE STATUS_CODE
    cont>     WHEN '1' THEN 'Full time'
    cont>     WHEN '2' THEN 'Part time'
    cont>     WHEN NULL THEN 'Unknown'
    cont>     ELSE 'Inactive'
    cont>   END
    cont> FROM EMPLOYEES;
     LAST_NAME        FIRST_NAME
     Smith            Terry        Part time
     O'Sullivan       Rick         Full time
       .
       .
       .
     Sarkisian        Dean         Part time
     Stornelli        James        Full time
     Hall             Lawrence     Full time
     Mistretta        Kathleen     Full time
     James            Eric         Inactive
     MacDonald        Johanna      Full time
     Dement           Alvin        Full time
     Blount           Peter        Full time
     Herbener         James        Full time
     Ames             Louie        Full time
    100 rows selected

    When SQL encounters the first WHEN clause that matches the
    primary value expression following the CASE keyword, it evaluates
    the THEN clause. If no matching values are found, the ELSE clause
    is evaluated. If the ELSE clause is missing, NULL is the returned
    value. For example:

    SQL> SELECT PRODUCT_NAME,
    cont>   CASE
    cont>      WHEN QUANTITY <= 0 THEN 'On back order'
    cont>      WHEN QUANTITY > 0 THEN
    cont>         CAST(QUANTITY AS VARCHAR(10)) || ' in stock'
    cont>   END
    cont> FROM INVENTORY;
     PRODUCT_NAME
     Staples-boxes     20 in stock
     Staplers-each     3 in stock
     Tape-rolls        On back order
     Calendars-each    25 in stock
     Tape disp.-each   On back order
     Desk cleaner      NULL
    6 rows selected

    An example of the searched case expression follows:

    SQL> SELECT PRODUCT_NAME,
    cont>   CASE
    cont>      WHEN QUANTITY <= 0 THEN 'On back order'
    cont>      WHEN QUANTITY > 0 THEN
    cont>         CAST(QUANTITY AS VARCHAR(10)) || ' in stock'
    cont>      ELSE                                          -- must be NULL
    cont>        'New Item - awaiting stock'
    cont>   END
    cont> FROM INVENTORY;
     PRODUCT_NAME
     Staples-boxes     20 in stock
     Staplers-each     3 in stock
     Tape-rolls        On back order
     Calendars-each    25 in stock
     Tape disp.-each   On back order
     Desk cleaner      New Item - awaiting stock
    6 rows selected

    The searched case expression allows arbitrary expressions in each
    WHEN clause, as shown in the previous example. The simple case
    expression is a shorthand method of specifying the searched case
    expression.

    For the simple and searched case expressions, the data types of
    the value expressions of the WHEN clause must be comparable, and
    the data types of the value expressions of the THEN clause must
    be comparable.

    All subqueries in a CASE expression are evaluated. It is the
    results of these subqueries that are conditionalized by the CASE
    expression and not the actual evaluation.

    If any subquery (which must return at most a single row and
    column) returns more than one row, the following exception is
    generated:

    %RDB-E-MULTIPLE_MATCH, record selection criteria should identify only one
    record; more than one record found

    A workaround is to add one of the following clauses to the
    subquery:

    o  LIMIT TO 1 ROW

       This ensures that only one row is returned. For example:

          .
          .
          .
       cont> WHEN A IS NOT NULL
       cont> THEN (SELECT A FROM T WHERE B = Y
       cont>       LIMIT TO 1 ROW)
          .
          .
          .

       The WHEN condition ignores this row if it is not valid.

    o  Duplicate the WHEN clause Boolean inside the subquery
       predicate

       For example:

       SQL> --
       SQL> -- Change the following syntax from
       SQL> --
          .
          .
          .
       cont> WHEN A IS NOT NULL
       cont> THEN (SELECT A FROM T WHERE B = Y)
          .
          .
          .
       SQL> --
       SQL> -- to include the Boolean inside the subquery
       SQL> --
          .
          .
          .
       cont> WHEN A IS NOT NULL
       cont> THEN (SELECT A FROM T WHERE B = Y AND A IS NOT NULL)
          .
          .
          .

       In this example, when the WHEN clause evaluates as FALSE, so
       will the WHERE predicate from the subquery and, therefore,
       will return no rows.

    In either of the above cases, the correct results are returned
    from the query.

4  –  DECODE

    The DECODE function compares an expression to each supplied
    search value until a match is found. When a match is found,
    DECODE returns the result in the corresponding result field. If
    no match is found, DECODE returns the default if it is specified,
    null if no default is specified.

    Example: Using the DECODE function

    SQL> SELECT employee_id, last_name, first_name,
    cont> DECODE (status_code, '1', 'Full time',
    cont>                      '2', 'Part time')
    cont> FROM employees
    cont> LIMIT TO 5 ROWS;
     EMPLOYEE_ID   LAST_NAME       FIRST_NAME
     00165         Smith           Terry        Part time
     00190         O'Sullivan      Rick         Full time
     00187         Lasch           Stan         Full time
     00169         Gray            Susan        Full time
     00176         Hastings        Norman       Full time
    5 rows selected

5  –  GREATEST and LEAST Functions

    The GREATEST and LEAST functions accept a list of two or more
    value expressions (all of which must be of comparable types) and
    return the greatest value from the list for the GREATEST function
    and the least value from the list for the LEAST function.
    The value expressions specified can be column references,
    subselects, function calls, literal values, and other complex
    value expressions.

    The data type of the resulting expression is a common data type
    to which all value expressions in the list can be converted. For
    example, LEAST(10, 10.3, 123E100) results in a DOUBLE PRECISION
    result because at least one literal is DOUBLE PRECISION.

    If the result data type resolves to a fixed CHARACTER string,
    then GREATEST and LEAST return a CHARACTER VARYING (also known as
    VARCHAR) string with the maximum length.

    The NULL keyword can appear in the list but is ignored. However,
    not all value expressions can be specified as NULL. That is, a
    non-NULL value expression must be in the list so that the data
    type for the expression can be determined.

    The GREATEST and LEAST functions can result in NULL only if at
    run time all value expressions result in NULL.

    If DATE VMS literals are used as an argument to the GREATEST or
    LEAST function, the date string must be prefixed with the type
    DATE VMS so that SQL will accept it as a DATE VMS literal and not
    a string literal. See the following examples.

    Examples: Using the GREATEST and LEAST functions

    Example 1 Using CHARACTER Versus DATE VMS Literals

    SQL> -- Different results are returned by the LEAST function (and
    SQL> -- the GREATEST function) if the parameters are treated as
    SQL> -- CHARACTER or DATE VMS literals.  This is because it is
    SQL> -- the data types of the source expressions that determine the
    SQL> -- the result data type of the LEAST (and GREATEST) functions.
    SQL> select LEAST ('1-APR-2001', '10-JAN-2000'),
    cont> LEAST (DATE VMS'1-APR-2001', DATE VMS'10-JAN-2000')
    cont> from rdb$database;

     1-APR-2001    10-JAN-2000 00:00:00.00
    1 row selected

    Example 2 finds the least value for the birthday of employees
    from two employees tables.

    Example 2 Using the LEAST Function

    SQL> SELECT LEAST (M.BIRTHDAY, E.BIRTHDAY, :SUPPLIED_DATE)
    cont> FROM EMPLOYEES E, JOB_HISTORY JH, DEPARTMENTS D, EMPLOYEES M
    cont> WHERE E.EMPLOYEE_ID = JH.EMPLOYEE_ID AND
    cont> ...

6  –  NULLIF Expressions

    The NULLIF expression is used to substitute NULL when two value
    expressions are equal. For example, if the data stored in column
    ADDRESS_DATA_1 or ADDRESS_DATA_2 are space characters, the NULLIF
    expression replaces the space value with the NULL value.

    SQL> BEGIN
    cont>   INSERT INTO EMPLOYEES
    cont>     (EMPLOYEE_ID, LAST_NAME, FIRST_NAME,
    cont>      ADDRESS_DATA_1, ADDRESS_DATA_2)
    cont>   VALUES
    cont>     (:EMP_ID, 'Clinton', 'William',
    cont>      NULLIF(:ADD_1, ' '),
    cont>      NULLIF(:ADD_2, ' '));
    cont> END;
    SQL>
    SQL> SELECT LAST_NAME, ADDRESS_DATA_1, ADDRESS_DATA_2
    cont> FROM EMPLOYEES
    cont> WHERE EMPLOYEE_ID = :EMP_ID;
     LAST_NAME        ADDRESS_DATA_1              ADDRESS_DATA_2
     Clinton          NULL                        NULL
    1 row selected

    The following example substitutes NULL when the MIDDLE_INITIAL
    column of the EMPLOYEES table contains space characters:

    SQL> SELECT LAST_NAME,
    cont>   NULLIF (MIDDLE_INITIAL, ' '),
    cont>   FIRST_NAME
    cont> FROM EMPLOYEES
    cont> WHERE EMPLOYEE_ID IN ('00191', '00198');
     LAST_NAME               FIRST_NAME
     Pfeiffer         I      Karen
     Gehr             NULL   Leslie
    2 rows selected

7  –  NVL Expressions

    substitutes NULL when two value expressions are equal; otherwise,
    returns NULL. The NVL expression is a synonym for the COALESCE
    expression. See the COALESCE_Expressions HELP topic for more
    information.

8  –  NVL2 Expressions

    NVL2 lets you compute a value based on whether a specified
    expression is null or not null. If the first value expression
    is not null then the second value expression is returned as
    the function result. Otherwise, the final value expression is
    returned. The data type function is derived as a common data type
    of the second and third value expressions.

    For example, when the JOB_END date in JOB_HISTORY is NULL then
    that indicates the current job for that employee. The following
    example uses NVL2 to annotate the output from a query on JOB_
    HISTORY displaying either "current job" or "prior job" based on
    the NULL attribute of the JOB_END column.

    SQL> select employee_id, job_start, job_end,
    cont> NVL2 (job_end, 'prior job', 'current job')
    cont> from job_history
    cont> where employee_id < '00180'
    cont> order by employee_id, job_start;
     EMPLOYEE_ID   JOB_START     JOB_END
     00164          5-Jul-1980   20-Sep-1981   prior job
     00164         21-Sep-1981   NULL          current job
     00165          1-Jul-1975    4-Sep-1977   prior job
     00165          5-Sep-1977    7-Apr-1979   prior job
     00165          8-Apr-1979    7-Mar-1981   prior job
     00165          8-Mar-1981   NULL          current job
       .
       .
       .

    The following example shows whether the income of some employees
    is made up of SALARY plus COMMISSION, or just SALARY, depending
    on whether the COMMISSION_PCT column of EMPLOYEES is null or not.

    SQL> SELECT last_name, salary_amount,
    cont>       NVL2 (commission_pct,
    cont>             salary_amount + (salary_amount * commission_pct),
    cont>             salary_amount) as Income edit using SALARY
    cont> FROM employees e, salary_history sh
    cont> WHERE last_name like 'B%'
    cont>   and e.employee_id = sh.employee_id
    cont>   and salary_end is null
    cont> ORDER BY last_name;
     E.LAST_NAME      SH.SALARY_AMOUNT        INCOME
     Babbin                 $20,150.00    $20,956.00
     Bartlett               $14,817.00    $15,261.51
     Bartlett               $38,223.00    $38,987.46
     Belliveau              $54,649.00    $55,741.98
     Blount                 $63,080.00    $64,341.60
     Boyd                   $30,275.00    $30,275.00
     Boyd                   $24,166.00    $24,166.00
     Brown                  $50,357.00    $50,357.00
     Burton                 $23,053.00    $23,053.00
    9 rows selected
    SQL>

9  –  SIGN Function

    SIGN returns an INTEGER value. SIGN accepts any numeric (fixed or
    floating) or interval value expression. If the value expression
    evaluates to NULL, then the SIGN function returns NULL. If the
    value expression evaluates to a negative value, then SIGN returns
    -1; if the value is positive then SIGN returns 1; otherwise a
    zero will be returned.

    Example: Using SIGN Builtin Function

    This example computes delayed departures from the LAYOVER table.

    SQL> select arr_date,
    cont>       dep_date,
    cont>       DECODE (SIGN ((dep_date - arr_date) day(9)),
    cont>           -1, 'date error - can not depart before arrival',
    cont>            0, 'same day departure',
    cont>            1, 'delayed')
    cont> from LAYOVER;

     2005-12-22   2006-01-20   delayed
     2005-12-23   2005-12-25   delayed
     2006-01-30   2006-02-01   delayed
     2006-02-06   2006-02-09   delayed
     2006-01-24   2006-01-26   delayed
     2006-02-02   2006-02-19   delayed
     2007-02-10   2007-02-16   delayed
     2007-02-20   2007-02-26   delayed
     2007-05-29   2007-06-08   delayed
     2007-06-12   2007-06-26   delayed
     2007-05-15   2007-05-21   delayed
     2007-09-10   2007-09-14   delayed
     2007-09-04   2007-09-06   delayed
     2007-09-19   2007-09-20   delayed
     2007-09-21   2007-09-24   delayed
    15 rows selected
Close Help