Numeric Input and Exponents

Go Up to Fixed-decimal Data Types

Any numeric string in DSQL or isql that can be stored as a DECIMAL(18,S) is evaluated exactly, without the loss of precision that might result from intermediate storage as a DOUBLE. A numeric string is recognized by the DSQL parser as a floating-point value only if it contains an “e” or “E” followed by an exponent, which may be zero. For example, DSQL recognizes 4.21 as a scaled exact integer, and passes it to the engine in that form. On the other hand, DSQL recognizes 4.21E0 as a floating-point value.