The Power Bi median does not return the right value.

[Sding]

This is the chart for the Great Sydney area's median price(the blue line) and the number of sold (the bar chart) for each year.  The shared x-axis is the year number, a generated column from date of each row (= Table.AddColumn(#"Filtered Rows1", "Year", each Date.Year([sold_contract_date]), Int64.Type)). The data is from a government agency. It seems impossible the median prices are so "rounded"(900k,950k,1000k,1000k).

So I filtered the year 2019, used the "matrix" chart with the same other conditions, and export the one-column table (a .csv file) out. Then open with Excel, It has 4453 rows, and the median (2287th row) is 118300.

I googled and read some posts from community.powerbi.com, but had no luck to find a similar question. I use DAX median function to do the same logic, which returns the same result. I just could not see any problems. The logic is very simple. There is only one filter of the other column from the same table. 

Many thanks to all in advance!

[Greg_Deckler]

@Sding - This will be very difficult to troubleshoot without the data, can you share it?

[Sding]

sure:

https://drive.google.com/file/d/1baDL95VTroXzAjv1KzADbtXCYv7O7CBP/view?usp=sharing

[Greg_Deckler]

@Sding - So, I exported out the main table to Excel. If I take the median of the entire sold_price column for 1,048,576 rows in Excel I get 584,000. In Power BI, I get 529,950 for all rows (1,890,467 rows). Interestingly, if you remove your page filter, you get less rounded values, so not quite sure why that is.

[Sding]

Hi Greg, thanks for your time and efforts.