Definition Attributes
Data Definition Language (DDL) Reference Manual — 426798-002
6- 48
TYPE Clause
•
5,000 bytes for an entire TACL structure, which allows 5,000 ASCII
characters or 2,500 national characters
•
Available address space for COBOL and TAL or, for part of a record, the
record length
BINARY
represents a two’s complement binary number, whose size can be specified as
8, 16, 32, or 64 bits.
scale
for COBOL, pTAL, and TAL only, is an integer that specifies the position of
an implied decimal point. A scale of
n
multiplies the number by 10 to the
power of
-n
; a scale of
-n
multiplies the number by 10 to the power of
n
.
The value of
scale
depends on the BINARY item size:
BINARY 16 scale
can be -4 to 4
BINARY 32 scale
can be -9 to 9
BINARY 64 scale
can be -18 to 18
(for COBOL, -17 to 18)
UNSIGNED
declares an item of type BINARY as a positive binary integer. BINARY 64
numbers must not be described as UNSIGNED.
FLOAT
for C, FORTRAN, pTAL, TAL, and Pascal only, represents a signed real
number in binary scientific notation. A FLOAT 32 field is 4 bytes long; a FLOAT
64 field is 8 bytes long. If the length is omitted,
data-type
is FLOAT 32.
COMPLEX
for FORTRAN and C only, represents an 8-byte binary complex number.
LOGICAL
for C, FORTRAN, Pascal, pTAL, TACL, and TAL only, represents a logical data
type. The value of a logical data type is considered false if it is 0 and true if it is
nonzero. A logical item can be either 2 bytes long (LOGICAL 2) or 4 bytes long
(LOGICAL 4). If the size is omitted,
data-type
is LOGICAL 2. LOGICAL 1 is
valid in DDL, C, and Pascal, but invalid in FORTRAN.
ENUM
represents an enumeration data type. For COBOL, Pascal, pTAL, and TAL,
DDL translates ENUM into a numeric data type. For C, DDL translates ENUM
data types into separate enumeration declarations with a short data type.
FORTRAN does not support ENUM.