DOUBLE, DOUBLE PRECISION
DOUBLE[(width,decimals)] [UNSIGNED] [ZEROFILL]
Use this data column type for approximate, floating-point
numbers. The width given is the maximum width of the total digits—on
both sides of the decimal place, not including the decimal point. The
decimals given are the number of decimals allowed and the number for
which it zero-fills. For example, a column type setting of
DOUBLE(4,2) that is given a value of 123.4 will
return a warning and store the value as 99.99, the maximum amount
given the width. If the same column is given a value of 12.3, it won’t
generate any warnings and it will store the value as 12.30.
The signed range of values can be from –1.7976931348623157E+308 to –2.2250738585072014E–308; the unsigned range from 2.2250738585072014E–308 to 1.7976931348623157E+308. Usually these limits aren’t possible due to the hardware and operating system limits of the server. The accuracy of this data type is up to about 15 decimal places.
With this data type, if UNSIGNED is
specified, negative values are not allowed. If the SQL mode for the
server does not have the REAL_AS_FLOAT option
enabled, REAL is a synonym for
DOUBLE.