Forum Discussion
MdxScript(Model)(17,3) calculation error
Yes, Int64 is the limit. Embarrassingly Power BI does not support Int128.
Have you tried the logarithm approach?
Super UserHello AnnieV,
Can you please try the following:
GLOS =
EXP (
AVERAGEX (
FILTER (
CAC_AllData,
NOT(ISBLANK(CAC_AllData[YourNumericColumn])) && CAC_AllData[YourNumericColumn] > 0
),
LN(CAC_AllData[YourNumericColumn])
)
)
I'm not sure I understand what your formula is doing. I need to calculate the product of all values in the column, then take the nth root of that product, where n is the number of rows in the column. They are all greater than 0 and not blank.
- lbendlin
RANDBETWEEN(1,1000)/100That can result in numbers greater than 1, worst case for all 10000 IDs. That can easily exceed the Int64 limits of Power BI when you multiply them. Change your calculation so that they all are below 1 (or use logarithms).
Many to most of the values will be over 1, as this is a calculation using a measure of time in days (datediff on hours, divided by 24 to get a decimal day value), with typical values between 1 and 3, but sometimes have much longer outliers.
Is Int64 the largest number it can deal with? I can't pre-calculate in a practical view in a SQL query, as the aggregation is dynamic (individual bucket, group of buckets, all my buckets).
I've thought of running an R script and pulling it back, storing as text if needed, but I'm not the Power BI admin, and that will be a slog that I need at least a temporary work around for.
