A floating-point number.
p represents the precision in bits.
If p is from 0 to 24, the data type becomes FLOAT with no M or D values.
If p is from 25 to 53, the data type becomes DOUBLE with no M or D values.