Hive uses a maximum or minimum value for integer and bigint data when there is data overflow during data type conversion.

If an expression results in numerical errors, such as division by zero or SQRT of a negative number, it returns an infinite or an NaN value. In the native environment, the expression returns null values and the rows do not appear in the output.

The results of arithmetic operations on floating point types, such as Decimal, can vary up to 0.1 percent between the native environment and a Hadoop environment.

You can use high precision Decimal data type with Hive 0.11 and above. When you run mappings on the Hive engine, the Data Integration Service converts decimal values with a precision greater than 38 digits to double values. When you run mappings that do not have high precision enabled, the Data Integration Service converts decimal values to double values.

When the Data Integration Service converts a decimal with a precision of 10 and a scale of 3 to a string data type and writes to a flat file target, the results can differ between the native environment and a Hadoop environment. For example, on the Hive engine, HDFS writes the output string for the decimal 19711025 with a precision of 10 and a scale of 3 as 1971. The Data Integration Service sends the output string for the decimal 19711025 with a precision of 10 and a scale of 3 as 1971.000.

The results of arithmetic operations on floating point types, such as a Double, can vary up to 0.1 percent between the Data Integration Service and the Hive engine.

When you run a mapping with a Hive target that uses the Double data type, the Data Integration Service processes the double data up to 17 digits after the decimal point.