Reference Guide > Function Support for Data Sources > Function Support Issues when Combining Data Sources > Data Precision
 
Data Precision
FLOAT and REAL Precision
Many data sources treat FLOAT and REAL as single-precision, but TDV treats these data types as DOUBLE. Queries can therefore return different results (more or fewer significant digits) depending on whether they are pushed or not pushed.
INTEGER Precision
When an value of INTEGER type is divided by another value of INTEGER type, the result might be INTEGER or it might be some other SQL Standard exact numeric type with implementation-defined precision and scale. So, for example, dividing 10 by 3 might produce exactly 3, or it might produce 3.3333.