#49 – When to Use the Decimal Type

You should use the decimal type (System.Decimal) for monetary calculations, or whenever you want more digits of precision, to avoid round-off error.

For example:

 float f1 = 12345678901234567.28f;     // 1.23456784E+16
 double d1 = 12345678901234567.28d;    // 12345678901234568.0
 decimal dc1 = 12345678901234567.28m;  // 12345678901234567.28

The decimal type gives you greater precision, but doesn’t support storing numbers as large as float or double.  This is because it stores all digits, rather than storing the mantissa and exponent of the number.  The decimal type also requires more storage space.

float – 4 bytes  (±1.5e−45 to ±3.4e38, 7 digit precision)
double – 8 bytes  (±5.0e−324 to ±1.7e308, 15-16 digit precision)
decimal – 16 bytes  (±1.0 × 10−28 to ±7.9 × 1028, 28-29 digit precision)