Forum Discussion
Snowflake Connector Fails in Power Query / Does Work in Workplace Connection Settings
Hello,
When I am editing my snowflake query in power query (service), I am getting the error "We could not evaluate this query due to invalid or missing credentials." I put in the credentials I have but it doesn't work. I'm using Key/Pair authentication.
At the same time, in my workplace "Manage Connections and Gateways" -> "Connections" the snowflake connection is showing as Online and checking status continues to show Online. If the key had rotated I would expect that to fail as well.
Why the discrepency?
I know my snowflake table has changed and I need to rename some fields but I don't get why that would cause a credential error. Outside of power query when I do a manual data refresh of my semantic model I am getting a field mismatch (known), (and not a credential problem) but I am running into issues on how to change it since I need to be able to connect to save changes to the M code for renaming.
Also - the connection is a direct cloud connection so it doesn't go through the on-prem data gateway.
Thanks,
Justin
Hi justincarter ,
Thank you for reaching out to the Microsoft Community Forum. You are facing a mismatch between “connection-level credentials” and “dataset/query editor credentials” in Power BI / Fabric with Snowflake + key pair auth + Direct cloud connections. It is not a credential problem It’s a schema mismatch causing query evaluation failure, which Power Query incorrectly reports as a credential error.
Please try below general workarounds to fix the issue.
1. Temporarily bypass schema validation: Open Power Query --> Go to Advanced Editor, modify your query to avoid immediate column dependency. Please refer below sample code.
let
Source = Snowflake.Databases(...),
Table = Source{[Name="Table"]}[Data],
#"Keep Columns Safe" = try Table otherwise Table
in
#"Keep Columns Safe"Then Save, Re-open and fix column names step-by-step.
2. If you know the column changes, Open Advanced Editor --> Update renamed fields manually in M code: Table.RenameColumns and Table.SelectColumns. Please refer below sample code.
Table.RenameColumns(Source, {
{"old_name", "new_name"}
})3. Go to: Dataset settings --> Data source credentials. Click Edit credentials, Re-enter key pair auth details and Save.
4. Duplicate the query (M code), Delete original Snowflake source step, Reconnect fresh and Reapply transformations.
I hope this information helps. Please do let us know if you have any further queries.
Regards,
Dinesh
2 Replies
Hi justincarter ,
Thank you for reaching out to the Microsoft Community Forum. You are facing a mismatch between “connection-level credentials” and “dataset/query editor credentials” in Power BI / Fabric with Snowflake + key pair auth + Direct cloud connections. It is not a credential problem It’s a schema mismatch causing query evaluation failure, which Power Query incorrectly reports as a credential error.
Please try below general workarounds to fix the issue.
1. Temporarily bypass schema validation: Open Power Query --> Go to Advanced Editor, modify your query to avoid immediate column dependency. Please refer below sample code.
let
Source = Snowflake.Databases(...),
Table = Source{[Name="Table"]}[Data],
#"Keep Columns Safe" = try Table otherwise Table
in
#"Keep Columns Safe"Then Save, Re-open and fix column names step-by-step.
2. If you know the column changes, Open Advanced Editor --> Update renamed fields manually in M code: Table.RenameColumns and Table.SelectColumns. Please refer below sample code.
Table.RenameColumns(Source, {
{"old_name", "new_name"}
})3. Go to: Dataset settings --> Data source credentials. Click Edit credentials, Re-enter key pair auth details and Save.
4. Duplicate the query (M code), Delete original Snowflake source step, Reconnect fresh and Reapply transformations.
I hope this information helps. Please do let us know if you have any further queries.
Regards,
Dinesh
Thanks for your suggestions Dinesh I used this as part of my solution. In addition to schema issues I learned my snowflake connection was only allowed from certain IP addresses so that was also part of my problem. Thanks again for your response.
