Implement float(alias: REAL) dataType

[acquamarin]

FLOAT (single precision floating-point number (4 bytes))

1. Grammar rule for defining FLOAT column:

To create a column with FLOAT datatype, simply declare the column as: COLUMN_NAME FLOAT

2. Adds aggregation/order by support for FLOAT.

3. Refactors the arithmetic operation framework:

We only define binary operations with both parameters having the same type.
For example:
We only define 3 operations for MOD operation:
INT64 % INT64,DOUBLE % DOUBLE,FLOAT % FLOAT
For other combinations of parameter types, we do casting.
For example: If we want to compute DOUBLE % INT64, we will convert INT64 to DOUBLE.

4. Implements to_float() function (converts an INT64 val to FLOAT).