In an ANSI-compliant database, columns declared as a DECIMAL(p) data type have a precision of p and a scale of zero, so that only integer values are stored. In a database that is not ANSI-compliant, DECIMAL(p) is a floating-point data type whose scale is large enough to store the exponential notation for a value.
For example, the following calculation shows how many bytes a DECIMAL(5) column requires in the default locale (where the decimal point occupies a single byte):
| 1 byte for the sign of the data value
1 byte for the 1st digit 1 byte for the decimal point 4 bytes for the rest of the digits (precision of 5 - 1) 1 byte for the e symbol 1 byte for the sign of the exponent 3 bytes for the exponent ------------------------------------------------- 12 bytes total |
Thus, "12345" in a DECIMAL(5) column is displayed as "12345.00000" (that is, with a scale of 6) in a database that is not ANSI-compliant.
Home | [ Top of Page | Previous Page | Next Page | Contents | Index ]