The BINARY type cannot be used in any of the following items:

• An index definition
• An external reference column

In addition, the following table lists items for which your ability to specify the BINARY type changes depending on its defined length.

Table 1-8 Items in which specifiable/not specifiable changes in a BINARY type defined length

Item	Defined length
	1 to 255	256 to 32,000	32,001 to 2,147,483,647
Sorting	Y	Y	N
Grouping	Y	Y	N
Set function	Y	Y	N
Search condition	Y	Y	N
Data insertion/updating#1	Y	Y	Y
Duplicate elimination	Y	Y	N
Set operations	Y	Y	N
Concatenation operation	Y	Y	Y#2
Scalar functions	Y	Y	Y#3
CASE expression	Y	Y	N
CAST specification	Y	Y	N

Legend:

Y: Can be used.

N: Cannot be used.

#1: When specifying a literal in an update or insertion value for a BINARY-type column, you can specify only a hexadecimal character string literal.

#2: BINARY-type concatenation operations with a maximum length of greater than or equal to 32,001 bytes can be used only on update values for the SET clause of the UPDATE statement.

#3: The only scalar functions for which a BINARY type with a defined length of 32,001 bytes or greater can be specified are the LENGTH, SUBSTR, and POSITION functions; on other scalar functions, the BINARY type cannot be used with a defined length of 32,001 bytes or greater.

When handling the BINARY type in a UAP, you can use a locator to allow you to process the SQL that handles BINARY type data without storing the data itself on a client. For a description of locators, see 1.13 Locator.