Data Type Comparison Rules

Numeric Values

A larger value is considered greater than a smaller one. All negative numbers are less than zero and all positive numbers. Thus, -1 is less than 100; -100 is less than -1.

The floating-point value NaN (not a number) is greater than any other numeric value and is equal to itself.

Date Values

A later date is considered greater than an earlier one. For example, the date equivalent of '29-MAR-2005' is less than that of '05-JAN-2006' and '05-JAN-2006 1:35pm' is greater than '05-JAN-2005 10:09am'.

Character Values

Character values are compared on the basis of two measures:

• Binary or linguistic sorting

• Blank-padded or nonpadded comparison semantics

The following subsections describe the two measures.

Binary and Linguistic Comparisons

In binary comparison, which is the default, Oracle compares character strings according to the concatenated value of the numeric codes of the characters in the database character set. One character is greater than another if it has a greater numeric value than the other in the character set. Oracle considers blanks to be less than any character, which is true in most character sets.

These are some common character sets:

• 7-bit ASCII (American Standard Code for Information Interchange)

• EBCDIC Code (Extended Binary Coded Decimal Interchange Code)

• ISO 8859/1 (International Organization for Standardization)

• JEUC Japan Extended UNIX

Linguistic comparison is useful if the binary sequence of numeric codes does not match the linguistic sequence of the characters you are comparing. Linguistic comparison is used if the NLS_SORT parameter has a setting other than BINARY and the NLS_COMP parameter is set to LINGUISTIC. In linguistic sorting, all SQL sorting and comparison are based on the linguistic rule specified by NLS_SORT.

Blank-Padded and Nonpadded Comparison Semantics

With blank-padded semantics, if the two values have different lengths, then Oracle first adds blanks to the end of the shorter one so their lengths are equal. Oracle then compares the values character by character up to the first character that differs. The value with the greater character in the first differing position is considered greater. If two values have no differing characters, then they are considered equal. This rule means that two values are equal if they differ only in the number of trailing blanks. Oracle uses blank-padded comparison semantics only when both values in the comparison are either expressions of data type CHAR, NCHAR, text literals, or values returned by the USER function.

With nonpadded semantics, Oracle compares two values character by character up to the first character that differs. The value with the greater character in that position is considered greater. If two values of different length are identical up to the end of the shorter one, then the longer value is considered greater. If two values of equal length have no differing characters, then the values are considered equal. Oracle uses nonpadded comparison semantics whenever one or both values in the comparison have the data type VARCHAR2 or NVARCHAR2.

The results of comparing two character values using different comparison semantics may vary. The table that follows shows the results of comparing five pairs of character values using each comparison semantic. Usually, the results of blank-padded and nonpadded comparisons are the same. The last comparison in the table illustrates the differences between the blank-padded and nonpadded comparison semantics.

Portions of the ASCII and EBCDIC character sets appear in Table 3-8 and Table 3-9. Uppercase and lowercase letters are not equivalent. The numeric values for the characters of a character set may not match the linguistic sequence for a particular language.

Object Values

Object values are compared using one of two comparison functions: MAP and ORDER. Both functions compare object type instances, but they are quite different from one another. These functions must be specified as part of any object type that will be compared with other object types.

Varrays and Nested Tables

Comparison of nested tables is described in "Comparison Conditions".

Data Type Precedence

Oracle uses data type precedence to determine implicit data type conversion, which is discussed in the section that follows. Oracle data types take the following precedence:

• Datetime and interval data types

• BINARY_DOUBLE

• BINARY_FLOAT

• NUMBER

• Character data types

• All other built-in data types

Data Conversion

Generally an expression cannot contain values of different data types. For example, an expression cannot multiply 5 by 10 and then add 'JAMES'. However, Oracle supports both implicit and explicit conversion of values from one data type to another.

SELECT salary + '10'
  FROM employees;

SELECT last_name
  FROM employees
  WHERE employee_id = '200';

SELECT last_name
  FROM employees 
  WHERE hire_date = '24-JUN-06';

Security Considerations for Data Conversion

When a datetime value is converted to text, either by implicit conversion or by explicit conversion that does not specify a format model, the format model is defined by one of the globalization session parameters. Depending on the source data type, the parameter name is NLS_DATE_FORMAT, NLS_TIMESTAMP_FORMAT, or NLS_TIMESTAMP_TZ_FORMAT. The values of these parameters can be specified in the client environment or in an ALTER SESSION statement.

The dependency of format models on session parameters can have a negative impact on database security when conversion without an explicit format model is applied to a datetime value that is being concatenated to text of a dynamic SQL statement. Dynamic SQL statements are those statements whose text is concatenated from fragments before being passed to a database for execution. Dynamic SQL is frequently associated with the built-in PL/SQL package DBMS_SQL or with the PL/SQL statement EXECUTE IMMEDIATE, but these are not the only places where dynamically constructed SQL text may be passed as argument. For example:

EXECUTE IMMEDIATE
'SELECT last_name FROM employees WHERE hire_date > ''' || start_date || '''';

where start_date has the data type DATE.

In the above example, the value of start_date is converted to text using a format model specified in the session parameter NLS_DATE_FORMAT. The result is concatenated into SQL text. A datetime format model can consist simply of literal text enclosed in double quotation marks. Therefore, any user who can explicitly set globalization parameters for a session can decide what text is produced by the above conversion. If the SQL statement is executed by a PL/SQL procedure, the procedure becomes vulnerable to SQL injection through the session parameter. If the procedure runs with definer's rights, with higher privileges than the session itself, the user can gain unauthorized access to sensitive data.

Implicit and explicit conversion for numeric values may also suffer from the analogous problem, as the conversion result may depend on the session parameter NLS_NUMERIC_CHARACTERS. This parameter defines the decimal and group separator characters. If the decimal separator is defined to be the quotation mark or the double quotation mark, some potential for SQL injection emerges.