Bug in Desktop, Version: 2.79.5768.721 64-bit (mars 2020) - Error when "Apply query changes"

[Anonymous]

Prewiew in Power Query works fine, but when when I try to update it fail with this error message:

Apply query changes"

Queries Metadata:

Failed to save modifications to the server. Error returned: 'An unexpected error occurred (file 'tmsavepointvalidation.cpp', line 458, function 'TMSavePoint::VerifyProxyCreationAllowed').

I have never seen this error before... May be it has to do with the new feature enhanced metadata format?

Reagards!

[Anonymous]

Hi,

"Does it only happened in the March 2020 version of Power BI Desktop?" Yes, as I said, when I downgraded to the feb-version it works fine. (I had to build the report from scratch, because of the changed format. The feb-version could not read the PBIX-file from the mach-version. And yes, I also tried to rebuild from scratch in the march-version and the error accoured again.

"Are there any change applied to power query editor before orrur this error?" Yes, the report worked in the beginning of developing the report. It's a report against a Postgres-database with hundres of schemas. The query uses a M-function and loops through the same table in differnt schemas to to build one big common table. Then I think the error accoured. Her is the code:

let
Source = (pSchema2 as text) => let
Source = PostgreSQL.Database("XXXXXXX", "insight"),
rm1000_f_application = Source{[Schema=pSchema2,Item="f_application"]}[Data],
#"Removed Other Columns" = Table.SelectColumns(rm1000_f_application,{"id", "date_applied", "date_deleted", "commission", "candidate", "gender", "source", "publishing_place_id", "age", "orgunit", "workarea", "workspecialization_id", "date_last_apply", "application_ref", "specific_ref"})
in
#"Removed Other Columns"
in
Source

And the calling code:

let
Source = #"Tenant filter Server2",
#"Invoked Custom Function" = Table.AddColumn(Source, "f_application2_", each fnc_application2([Schema])),
#"Removed Columns" = Table.RemoveColumns(#"Invoked Custom Function",{"id", "name", "Server", "Schema"}),
#"Expanded f_application2_" = Table.ExpandTableColumn(#"Removed Columns", "f_application2_", {"id", "date_applied", "date_deleted", "commission", "candidate", "gender", "source", "publishing_place_id", "age", "orgunit", "workarea", "workspecialization_id", "date_last_apply", "application_ref", "specific_ref"}, {"id", "date_applied", "date_deleted", "commission", "candidate", "gender", "source", "publishing_place_id", "age", "orgunit", "workarea", "workspecialization_id", "date_last_apply", "application_ref", "specific_ref"}),
#"Changed Type" = Table.TransformColumnTypes(#"Expanded f_application2_",{{"KeyConstant", Int64.Type}, {"id", Int64.Type}, {"date_applied", type text}, {"date_deleted", type text}, {"commission", Int64.Type}, {"candidate", Int64.Type}, {"gender", Int64.Type}, {"source", Int64.Type}, {"publishing_place_id", Int64.Type}, {"age", Int64.Type}, {"orgunit", Int64.Type}, {"workarea", Int64.Type}, {"workspecialization_id", Int64.Type}, {"date_last_apply", type text}, {"application_ref", Int64.Type}, {"specific_ref", Int64.Type}}),
#"Split Column by Positions" = Table.SplitColumn(#"Changed Type", "date_applied", Splitter.SplitTextByPositions({0, 4, 6}), {"date_applied.1", "date_applied.2", "date_applied.3"}),
#"Split Column by Positions1" = Table.SplitColumn(#"Split Column by Positions", "date_deleted", Splitter.SplitTextByPositions({0, 4, 6}), {"date_deleted.1", "date_deleted.2", "date_deleted.3"}),
#"Split Column by Positions2" = Table.SplitColumn(#"Split Column by Positions1", "date_last_apply", Splitter.SplitTextByPositions({0, 4, 6}), {"date_last_apply.1", "date_last_apply.2", "date_last_apply.3"}),
#"Merged Columns" = Table.CombineColumns(#"Split Column by Positions2",{"date_applied.1", "date_applied.2", "date_applied.3"},Combiner.CombineTextByDelimiter("-", QuoteStyle.None),"Date Applied"),
#"Merged Columns1" = Table.CombineColumns(#"Merged Columns",{"date_deleted.1", "date_deleted.2", "date_deleted.3"},Combiner.CombineTextByDelimiter("-", QuoteStyle.None),"Date Deleted"),
#"Merged Columns2" = Table.CombineColumns(#"Merged Columns1",{"date_last_apply.1", "date_last_apply.2", "date_last_apply.3"},Combiner.CombineTextByDelimiter("-", QuoteStyle.None),"Date Last Apply"),
#"Changed Type1" = Table.TransformColumnTypes(#"Merged Columns2",{{"Date Applied", type date}, {"Date Deleted", type date}, {"Date Last Apply", type date}}),
#"Replaced Errors" = Table.ReplaceErrorValues(#"Changed Type1", {{"Date Deleted", null}}),
#"Replaced Errors1" = Table.ReplaceErrorValues(#"Replaced Errors", {{"Date Last Apply", null}}),
#"Replaced Errors2" = Table.ReplaceErrorValues(#"Replaced Errors1", {{"Date Applied", null}})
in
#"Replaced Errors2"

"Does this error only happen in spetical report or all  reports? " I have seen the error only in this report.

"Have you tried to re-install the Power BI Desktop or use the MS Store version? " Yes, many times! (Switch between feb. and marcg-version.

Regards

Aslak

[v-lid-msft]

Hi @Anonymous ,

Thanks for your additional details, your queries seems very fine, could you please try to verify whichi step will cause this error by creating a backup first, then delete last one step in the steps panel and save each time? Ineed, there is  a new feature about metadata included in march 2020, but it did not enable defaultly, so we do not think it will be the reason of this issue.

Hi @Anonymous , @Anonymous ,

Do you connect a postgresql data source too?

Best regards,

[Anonymous]

Yes I do, running into more Postgre Specific issues as now as well. The new meta data set up does not seem to want to play nicely with redshift. 

Now I'm consistently getting invalid digit errors even when I completely remove the offending column, it's like the meta data is preserving that it was there or something along those lines. 

[Anonymous]

hi @v-lid-msft ,

in response to your question: no, am not using Postgres. This report is using mock data in CSV files. It will be deployed to connect to SQL Server when ready.

However, and noting the other points raised on this thread I have some similar factors:

• The CSV files are stored on OneDrive for Business, that has SSO for authentication using an organisation account
• I have deleted unwanted Query Folders
• I have updated to the latest PowerBI this week, and enabled Preview Features

Cheers

[Anonymous]

Have just turned OFF the Preview Feature "Store datasets using enhanced metadata format" and restarted PowerBI Desktop, and this has now resolved the issue for me. I can now make changes in Transform/PowerQuery and they apply and update the report fine.

[gooranga1]

@Anonymousgreat spot. Just done the same and my query now works.

[Anonymous]

I'm connecting to a redshift DB + a onedrive file that I merge to one of the redshift pulls for some hierarchy mapping.

Potential Issues:
1.) The only change I made on the query side was merging two tables from the same DB together. 

2.) I read elsewhere people were running into this issue when they delete query folders, which is something that I did do yesterday so I'm looking into that.

3.) I always enable all preview features, while reading up on the enhanced metadata refresh I learned that once converted you can no longer use SSO data sources. Would connecting to my onedrive file using my organizational account count as SSO?

[JoeRobert06]

i'm seeing the same issue. This report was working just fine a couple days ago and now when i go to refresh, i get the sqldumper.exe prompt and this error message.

March 2020 (2.79.5768.1082) 64-bit

Also to confirm, i disabled enhanced dataset metadata and am still getting the same error message.

[Anonymous]

I believe this error occurs when you delete folders with queries inside of them! 

[v-lid-msft]

Hi @Anonymous , @Anonymous , @Anonymous , @gooranga1 , @gooranga1 ,

Based on the test result of @Anonymous  & @gooranga1 ( thank you for both) ,  it seems like the root reason is the preview feature  enhanced dataset metadata  is enabled But it seems many kind of operations will result in this issue, the only steps we can reproduce is deleting a folder with query ( thank your for @Anonymous 's information) , we will report this issue and then update here if we got anything new,

Currently the workaround is creating a backup then disable the preview feature enhanced dataset metadata in options.

Best regards,

[Skwalla]

I'm struggling with the same error..

Yesterday I deleted query folders with unwanted queries, applied changes and it worked fine. Today I have just made a few changes to the query itself, and it gives me the error.

I'm working with excel and csv data sources.

The workaround with deactivating the preview feature does not work for me though..

[gooranga1]

I am getting the same error connecting to a public website.

let
    Source = Json.Document(Web.Contents("https://opendata.ecdc.europa.eu/covid19/casedistribution/json/")),
    #"Converted to Table" = Record.ToTable(Source),
    #"Expanded Value" = Table.ExpandListColumn(#"Converted to Table", "Value"),
    #"Expanded Value1" = Table.ExpandRecordColumn(#"Expanded Value", "Value", {"dateRep", "day", "month", "year", "cases", "deaths", "countriesAndTerritories", "geoId", "countryCode", "popData2018"}, {"Value.dateRep", "Value.day", "Value.month", "Value.year", "Value.cases", "Value.deaths", "Value.countriesAndTerritories", "Value.geoId", "Value.countryCode", "Value.popData2018"})
in
    #"Expanded Value1"

 Works ok in Power Query but errors when I Apply.

[Anonymous]

Running into the same issue, PBI is my full-time job and I've also never seen this error. 

[amitchandak]

Check and log a ticket on https://powerbi.microsoft.com/en-us/support/