Forum Discussion
Date format change when deploying through pipeline
Hey there,
i got an issue with several date columns in a power bi dataset. The format as set in desktop is dd.mm.yyyy and when i publish in dev workspace it is still the same. The problem occurs when i deploy to test workspace. It switches to mm/dd/yyyy which brakes the model which is heavily dependent on dynamic parameters.
Unfortunately it is impossible to post the model, since the problem is not reproducible for me.
Any help is appreciated.
Kind regards
Chris
Hi EasyChreasi ,
I have a couple of ideas/suggestions to look into.
Non-standard ISO dates (i.e. a date format other than yyyy/mm/dd) can be very unpredictable in Fabric/Power BI.
What is your source date format? Are they from a csv? It might be treating non-ISO formatted dates as strings when converting to a date in Power BI and is not treating the dates consistently across workspaces. Consider forcing an ISO date (yyyy/mm/dd) in Power Query and then setting the date to your preferred date format in desktop visual formatting.
Also another thing to look at, which might be a long shot, but is the locale/language the same between dev and test workspaces? If the locale settings are different that might cause the issue you are seeing.
Chris Andrews, Senior Data Analyst
www.linkedin.com/in/dr-christopher-andrews
9 Replies
Hi EasyChreasi ,
I have a couple of ideas/suggestions to look into.
Non-standard ISO dates (i.e. a date format other than yyyy/mm/dd) can be very unpredictable in Fabric/Power BI.
What is your source date format? Are they from a csv? It might be treating non-ISO formatted dates as strings when converting to a date in Power BI and is not treating the dates consistently across workspaces. Consider forcing an ISO date (yyyy/mm/dd) in Power Query and then setting the date to your preferred date format in desktop visual formatting.
Also another thing to look at, which might be a long shot, but is the locale/language the same between dev and test workspaces? If the locale settings are different that might cause the issue you are seeing.
Chris Andrews, Senior Data Analyst
www.linkedin.com/in/dr-christopher-andrewsHi EasyChreasi,
This is interesting, I've never seen a deployment pipeline change the content of a model like that.
What do you mean that you cannot reproduce this? If you run the deployment pipeline does it work as intended?
No, i mean i cannot reproduce this with another dataset.
No, i mean i cannot reproduce this with another dataset. Some additional information - We are using a direct query model - both workspaces point to the exact same database tables on a redshift.
Hi EasyChreasi,
If you're able to recreate the model and not have any issues with it, then there must be a problem somewhere with the configuration of your original model.
I'm not sure what that problem is, but I would probably just recreate the model and move on. As long as the issue doesn't come back, I'd mark it up as a fluke and not think too much harder about it.
You can also open a support ticket and see what the support engineers say about the issue.
it is a really complex model - several weeks of work - not solving this is not an option.
Update: The problem occurs, when i refresh the data in the dataset. The date tables are other than most tables in the dataset import mode.
Hi EasyChreasi,
Thank you for reaching out to Microsoft Fabric Community.
Thank you tayloramy and chris_andrews for the prompt response.
As we haven’t heard back from you, we wanted to kindly follow up to check if the response's provided by the user's for the issue worked? or let us know if you need any further assistance.
Thanks and regards,
Anjan Kumar Chippa
Hi @EasyChreasi,
We wanted to kindly follow up to check if the response's provided by the user's for the issue worked? or let us know if you need any further assistance.
Thanks and regards,
Anjan Kumar Chippa
