value-expression::=[sign]term[operator[sign]term]...
    term::=[cast-specification]value-expression-primary
    operator::={+|-|*|/}
    sign::={+|-}
    cast-specification::={'('TINYINT')'|'('SMALLINT')'|'('INT[EGER]')'
                   |'('BIGINT')'|'('REAL')'|'('FLOAT')'|'('DOUBLE')'
                   |'('DEC[IMAL]')'}
    value-expression-primary::={column-specification|column-name|constant|'('value-expression')'}

Specifies a value.

sign

Specify + or -.

term

Specify a value expression primary and, optionally, a cast specification.

cast specification

For details about the data types that can be specified for cast specification, see 3.3 CQL data types.

value expression primary

Specify a column specification, a column name, a constant, or a value expression.

column-specification

For details about column specifications, see 3.2.5(2) Column specification.

column-name

If the succeeding FROM clause only references a single relation, you can specify just a column name.

If the succeeding FROM clause references multiple relation, you must specify a column specification.

constant

For details about constants, see 4.4.20 Constant.

operator

Specify +, -, *, or /. When specifying a singed term after an operator, enclose the term in parentheses. Signs and operator examples follow:

-a+(-b)
a*(-b+1)
a/(-b)+1

• You can perform the four arithmetic operations by specifying them in a value expression.
• The data type of the result of a value expression is the same as the data type of the result of the four arithmetic operations or of a value expression primary. The data structure is also the same as the arithmetic operations or a value expression primary.
• The four arithmetic operations (binomial operations) only use numeric data. Therefore, non-numeric data (such as character data and time data) cannot be included in an operation. If it is included, an error occurs.
  Operations can be executed between numeric data of different types. The following table shows the relationship between the data types used in binomial operations and the resulting data type.
  

  Table 4-8 Relationship between the data types used in the four arithmetic operations (binomial operations) and the resulting data type

  No.	Data type of first operand	Data type of second operand
  		TNY	SML	INT	BIG	RL	FLT	DBL	DEC	NUM
  1	TNY	TNY	SML	INT	BIG	RL	FLT	DBL	DEC	NUM
  2	SML	SML	SML	INT	BIG	RL	FLT	DBL	DEC	NUM
  3	INT	INT	INT	INT	BIG	RL	FLT	DBL	DEC	NUM
  4	BIG	BIG	BIG	BIG	BIG	RL	FLT	DBL	DEC	NUM
  5	RL	RL	RL	RL	RL	RL	FLT	DBL	DEC	NUM
  6	FLT	FLT	FLT	FLT	FLT	FLT	FLT	FLT	DEC	NUM
  7	DBL	DBL	DBL	DBL	DBL	DBL	DBL	DBL	DEC	NUM
  8	DEC	DEC	DEC	DEC	DEC	DEC	DEC	DEC	DEC	DEC
  9	NUM	NUM	NUM	NUM	NUM	NUM	NUM	NUM	NUM	NUM

  Legend:

  TNY: TINYINT

  SML: SMALLINT

  INT: INTEGER

  BIG: BIGINT

  RL: REAL

  FLT: FLOAT

  DBL: DOUBLE

  DEC: DECIMAL

  NUM: NUMERIC

  
  
• For an operation that includes the DECIMAL or NUMERIC type, values in other data types are rounded according to the value of the property listed below and converted into values in the DECIMAL type (NUMERIC type) before they are used in the operation.
  
  • The query.decimalMaxPrecision property of system_config.properties
    This property specifies precision.
    
  • The query.decimalRoundingMode property of system_config.properties
    This property specifies a rounding method.
    
  The operation result is also rounded in the same way.
  

• If an overflow occurs when performing a binomial operation, processing continues. In this case, an integer will remain in the overflowed state (the higher bytes that do not fit are lost and the lower bytes are used as is) and a floating-point value will become infinity. If an overflow occurs in a floating-point value during an operation involving both a DECIMAL or NUMERIC type value and a floating-point value, however, an error (query group lockup) occurs.
• If 0 is specified for the second operand of a division operation, an error occurs.
• An binomial operation on character data or character string constant results in an error.
• Casting involving date/time data (including character string constant) results in an error.
• For character data, only character data whose format matches numeric data can be cast. For details about character data formats, see 3.4 Data comparison.
• When a value is cast into the DECIMAL type, it is rounded according to the value of the property listed below before being converted.
  
  • The query.decimalMaxPrecision property of system_config.properties
    This property specifies precision.
    
  • The query.decimalRoundingMode property of system_config.properties
    This property specifies a rounding method.
    

Outputs the values obtained by casting column a in relation s1 into DECIMAL data and column b into INT data. The underlined parts indicate value expressions.

REGISTER QUERY q1 SELECT (DECIMAL)s1.a AS xxx, (INT)s1.b AS yyy FROM s1[ROWS 100];