Power BI: Losing precision by using SUMMARIZECOLUMN?

Question

I am new to Power BI and I am making a matrix for Sales value. As we use various currencies and convert the numbers, each Sales number can have as many as 15 decimal points. NumericColumn corresponds to Sales going forward.

There is this OriginalTable:

NumericColumn,
OtherColumn1,
... ,
OtherColumnN

This OriginalTable is imported from SQL Server and has ~ 400K rows. NumericColumn has type numeric(35,15).

As there are a lot of unimportant columns in OriginalTable I made a new table in the following manner where M << N:

NewTable = SUMMARIZECOLUMNS(
OriginalTable[NumericColumn],
OriginalTable[OtherColumn1],
... , 
OriginalTable[OtherColumnM]
)

The sum of the "NumericColumn" in OriginalTable is 2,550,832.98 . This number is seen by putting "NumericColumn" in Value field as Sum and without setting any Rows or Columns in Visualization -> Matrix.

On the other hand, the sum of the "NumericColumn" in NewTable is 2,550,829.14 (3.84 smaller than that in OriginalTable).

It seems that I lost precision when I created a new table. Would somebody kindly take a look? Please let me know if more information is needed.

Thanks & Regards,

Kyoto

mkRabbani mkRabbani · Accepted Answer · 2020-10-12T06:39:20

By default, the M query transformation keeps up to 8 decimal points when you are importing data and applying Change Type. As you said you have up to 15 decimal points, you can try the below steps and check it worked for you or not.

Come back to report clicking Cloase & Apply button. Now in the report, select your decimal value column configure the column Format and Decimal points as shown below-

Now applying SUMMARIZECOLUMNS should give your expected result.

Power BI: Losing precision by using SUMMARIZECOLUMN?

1 Answers