Power BI Rounds Int64 columns to next thousand

[DmitryKo]

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?

[DmitryKo]

Any other insights on this? I'm rather surprised of this bug, to say the least

[Greg_Deckler]

@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.

[DmitryKo]

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.