Dataflow Gen2 Mashup Error

[RZidaj]

Error Code: Mashup Exception Expression Error, Error Details: Couldn't refresh the entity because of an issue with the mashup document MashupException.Error: Expression.Error: We cannot convert a value of type Table to type Text. Details: Reason = Expression.Error;ErrorCode = 10276;Value = #table({"Value"}, {});Microsoft.Data.Mashup.Error.Context

I'm creating a new dataflow and I keep getting hit with this error. the problem is I'm not sure which step is causing it because when I run this query in pbix it works fine but it doesn't work as a dataflow

I think the problem might be at this step but I don't know how to resolve it as I need the data to come through as a table under the Issues field so that I can expand it and view the columns

[stoic-harsh]

Hey @RZidaj,

If you want the 'Issues' column to contain a nested table that you can later expand, use a row-context filter explicitly.

Example:

Table.AddColumn(
    #"Filtered Rows",
    "Issues",
    each
        try
            Table.SelectRows(
                fyleDetails,
                (x) => x[Project Key] = [Project Key]
            )
        otherwise
            #table({}, {})
)

This returns a proper table per row. Then you can safely expand 'Issues'.

Your current expression is effectively returning:

• sometimes a List
• sometimes a Table
• sometimes null

Power BI Desktop tolerates this more than Dataflows do.

Dataflows require stronger type consistency during cloud refresh. In the updated expression, now the column is always a 'table' type.

Best,

Harshit

[RZidaj]

Thanks for explaining this to me clearly While your suggested query returned a different result than I was expecting (it returned only a single row of result as opposed to all rows) it helped me understand where my problem was

[v-achippa · Accepted solution]

Hi @RZidaj,

Thank you for reaching out to Microsoft Fabric Community.

Here looks like the issue is related to Dataflow Gen2 handling the Issues column as text during the refresh, while the function is actually returning a table.

So please try explicitly keeping the column type as table and avoid returning null. Please try this below query:

Table.AddColumn(#"Filtered Rows", "Issues", each try fnGetAllIssues([Project Key]) otherwise #table({}, {}), type table)

Also please check and remove any automatic changed type step for the Issues column if present.

Thanks and regards,

Anjan Kumar Chippa

[RZidaj]

Thanks so much Anjan!

Your suggested query returned the result (all rows). Unfortunately I pressed the solution button a bit too quick and I'm still getting the same mashup error when selecting save & run 

Error Code: Mashup Exception Expression Error, Error Details: Couldn't refresh the entity because of an issue with the mashup document MashupException.Error: Expression.Error: We cannot convert a value of type Table to type Text. Details: Reason = Expression.Error;ErrorCode = 10276;Value = #table({"Value"}, {});Microsoft.Data.Mashup.Error.Context =

[RZidaj]

I'm wondering if this could be the problem step

I've just updated it to this below but it's also still returning the same error

Table.FromList(Source, Splitter.SplitByNothing(), {"PRs"}, #table({}, {}), ExtraValues.Error)

[v-achippa]

Hi @RZidaj,

Here looks like the issue is still related to the PRs column. The Table.FromList step looks fine, please change the step back to:

Table.FromList(Source, Splitter.SplitByNothing(), {"PRs"}, null, ExtraValues.Error)

The error is likely happening in a later step during refresh. Please try removing any auto-generated Changed Type after the PR and Issues column is created and explicitly expand the columns directly after creation.

Dataflow Gen2 refresh is somewhat stricter than desktop when handling the nested record or table values, so queries that work in PBIX preview can still fail during Save & Run.

Thanks and regards,

Anjan Kumar Chippa

[v-achippa]

Hi @RZidaj,

As we haven’t heard back from you, we wanted to kindly follow up to check if the solution I have provided for the issue worked? or let us know if you need any further assistance.

Thanks and regards,

Anjan Kumar Chippa

[RZidaj]

Hi Anjan!

Just letting you know that I found the issue and unfortunately it was actually rooted in my function. Once I changed the otherwise null in the function too, that's when the issue resolved. Thank you so much for checking if your solution worked!

[lbendlin]

Your "otherwise" branch needs to return a table stub, not null.