Earn a 50% discount on the DP-600 certification exam by completing the Fabric 30 Days to Learn It challenge.
Simple table with EmployeeID column that has underlying Int64 data type. Source is import from Excel if it matters.
Problem: the model rounds the data value to the closest thousand. This is how it looks in Power BI Desktop (page configured with page-level filter based on 2 unique names so it's exactly the same 2 rows of data); changing display type to Text results in correct value being displayed; changing it back to "Whole number" results in incorrect rounded-to-thousand value displayed:
This is not a display-only issue as it results in incorrect data pulled through relationships. Note that the value itself (4 000 600 100 000 172 836) is within int64 range (int64.maxvalue = 9 223 372 036 854 775 807).
Why?
Any other insights on this? I'm rather surprised of this bug, to say the least
@DmitryKo You could try decimal number. You really want that as Text though. Using ID's like that as a number tends to remove leading zeros, etc. I highly recommend you leave it as Text.
The id in the datasource is int64, so leading zeroes are not an issue - if the data is interpreted correctly. Quesion is: why it's not being interpreted correctly by power bi (power query engine)?
And no, I don't want it as text; my assumption is performance on int64-based key vs text-based key would be orders of magnitude different.
P.S. Are you seriously suggesting changing ID column to a floating point data type?..
P.P.S. Please fix your signature, your 2 rows of text long reply takes almost entire screen due to clutter in the sig.
User | Count |
---|---|
98 | |
90 | |
82 | |
73 | |
67 |
User | Count |
---|---|
115 | |
102 | |
98 | |
71 | |
67 |