Calculating with High Precision
By setting a specific accuracy, you decide which number format is used internally for calculating key figures. The level of calculation accuracy can be specified in the InfoObject maintenance transaction for an individual key figure (system-wide), and specified in the query monitor for all key figures of a particular query.
Decimal floating point numbers are used for calculation in the analytic manager. This avoids problematic conversions of the data type float from the decimal system to the dual system and the other way round.
There are two levels of precision for decimal floating point numbers:
- Decimal floating point numbers with 16 places ("decfloat16", "short decfloat")
- Decimal floating point numbers with 34 places ("decfloat16", "long decfloat")
Short decfloats are used by default in the analytic manager for numerical data, to minimize the memory requirements for the main memory and the OLAP cache. If you choose Calculate with High Precision in the query monitor, long decfloats are used instead.
Choose this option for application areas where a high accuracy is required (for queries in the accounting and cost accounting areas for example).
The processing effort for long and short decfloats is similar. Due to the higher memory requirements for long decfloats, we recommend that you do not use this number format for all query key figures. Only use it for key figures that require a high degree of accuracy. You can do this by using the calculation precision setting for the required key figures in the InfoObject maintenance screen.
If you want the analytic manager to calculate all key figures of the selected query by using decimal floating point number with 34 places, set the relevant flag in the query properties:
- Deselect the InfoProvider Setting option.
- Choose Calculate with High Precision.