BUG: Double Precision Synonym Mapped to REAL Instead of FLOAT

SYMPTOMS

The "float and real (T-SQL)" topic in the SQL Server 7.0 Books Online defines the synonym for 'double precision' as float(53). The number 53 is the maximum number of bits used to store the mantissa of the number in scientific notation. The storage size for the float(53) data type is 8 bytes with a precision of 15 digits.

If a table is defined with a double precision column immediately following a character or varchar datatype, the 'double precision' synonym is mapped to a REAL data type. A REAL data type has precision of float(24), with a storage size of 4 bytes and a precision of 7 digits.

The following example demonstrates this behavior:

create table <tablename>
(
   strData      char(1),
   fData      double precision
)
go
				

This can be verified by executing either of the following statements:

   Exec Sp_help <tablename>
   Exec Sp_columns <tablename>