Can you post some sample data set. I guess in live you are not filtering teh data how you filtered it in extract. please check you are doing the same on both live and extract.
Hi, not sure if I can do that now as I'm pulling the entire data set. I'll see if I can minimize.
In terms of filtering, we haven't applied any, it's initially an extract connection switched to live, then we saw the discrepancy.
Please check your calculated filed which is applied in extract and live whether it is getting the right data or not.
1 of 1 people found this helpful
Sharing Tableau support's feedback and suggested workaround, for future reference:
Typically, the BigQuery documentation for Count Distinct states that the results are a statistical approximate. Given that these two calculations both use COUNTD() in their formulas, and conditions are different: COUNTD() in calculation 1 is applied to all Account_IDs, while COUNTD() in calculation 2 is just applied to Account_IDs with [New/Existing] = ‘New’, it could be possible that the results do not have the same deviation, thus the final results of these two calculations are not identical to each other.
When an extract is created, Tableau Desktop will give the exact count distinct of the extracted data, which is an accurate value, and the two calculations would have the same result. Because of this feature of BigQuery, the Count Distinct values may be different between extracts and live data when connecting to Google BigQuery.
While the following article is not supported by Technical Support it provides a workaround for this issue:
Tableau Tip: Count Distincts with Google BigQuery