Median calculation when using the Summarize tool is incorrect.

[jvolturo]

I've been using the summarize tool to calculate the median of stats by player by grouping on the player id and choosing Median and/or Percentile (50th), and I've the results are significantly incorrect.  If I isolate my dataset to one player at a time, the calculations are correct.  Has a solution to this error been solved yet?

[caltang]

What do you mean by incorrect? Do you have some sample data that can showcase this? That's quite a big finding if true.

[jvolturo]

Hello, thanks for the quick response.  So there are three ways that I'm validating the Median and Percentile calculations:

1. I've also calculated the Median and Percentiles in Power BI and the values are different,
2. I run the workflow with the summation tool calculating a Median, 10th Percentile, and 90th Percentile and the values change 1 out of every 5 times, even though the dataset is static,
3. if I calculate the Median in the summation tool for a single player, the value is correct, but when I try to calculate a median for all players simultaneously, the values change 1 out of every 5 times,

It's very weird, and makes me worried that my Standard Deviation, Variation, and other statistical calculations are also being miscalculated.  Luckily I'm only using NBA player data to calculate these values, but I also work in Healthcare and Technology consulting and I use these calculations for clients pretty often.

[caltang]

Since you're using NBA data, can you export your workflow with the data attached? I'd like to see it.