Schedule/manual refresh not possible from Azure Blob storage?

[jPinhao]

We have a blob storage account on Azure, from which we've built a Power BI model in the desktop app. It works fine on the desktop, and I can publish it to the online service without a problem. However when I try to schedule a refresh (or manually refresh) in the online service I get the following message:

"You can't schedule a refresh for this dataset because one or more sources currently don't support refresh"

I'm at a loss why it doesn't allow me to refresh it online. I'm sure I've done refreshes on models using Azure Blob storage before, and I'm not sure what could be different with this one. It doesn't even let me insert the required authentication data, which makes me think this sin't an issue with permissions either.

Has anyone had issues with connecting to Azure Blob storage before from the online service before? Any ideas why it might not be working?

[v-ljerr-msft]

Hi @jPinhao,

I have tested it on my side, and I can do a schedule/manual refresh for my dataset from Azure Blob Storage.

Is it possible that the dataset you are using contains other data sources that are not supported for refresh. Please go to Query Editor, Advanced Editor in Power BI Desktop to find the data source that you are using and compare that against the list of supported data sources at Refresh data in Power BI. Reference: Troubleshooting unsupported data source for refresh

Regards

[jPinhao]

Thanks for the reply @v-ljerr-msft. I had another look and it doesn't look like I have any other external sources setup.. It's the blob storage + a couple of tables which are hand created within Power BI. This is a screenshot of relevant section of the the dependencies diagram:

I also double checked the script for Load Data, and it seems pretty innocuous:

let
    Source = AzureStorage.Blobs(Text.Combine({"https://", StorageAccount, ".blob.core.windows.net/", StorageContainer})),
    #"Combined Binaries" = Binary.Combine(Source[Content]),
    #"Imported Text" = Table.FromColumns({Lines.FromBinary(#"Combined Binaries",null,null,1252)}),
    #"Parsed JSON" = Table.TransformColumns(#"Imported Text",{},Json.Document),
    #"Expanded Column1" = Table.ExpandRecordColumn(#"Parsed JSON", "Column1", {"event", "internal", "context"}, {"event", "internal", "context"})
in
    #"Expanded Column1"

Everything else within the model is either a function, or tables generated from "Load Data".

In the AzureStorage.Blobs line I use a couple of parameters defined in the pbix project - I don't suppose this would cause any issues with loading it in the service would it?

I also checked both hand-added tables and they do indeed only use a Table.FromRows and they both use:

Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("binaryText....", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type text) meta [Serialized.Text = true]) in type table [Database = _t, DB_Pictures_database = _t]),

As their source.. Again, I don't imagine this would cause any issues to the online service to load?

Any ideas?

[v-ljerr-msft · Accepted solution]

Hi @jPinhao,

---

In the AzureStorage.Blobs line I use a couple of parameters defined in the pbix project - I don't suppose this would cause any issues with loading it in the service would it?

---

Data sources dependent on parameters cannot be refreshed currently. That's because Power BI Service doing static analysis of the query to discover the data source, and static analysis can't yet handle this scenario. Here is a similar thread for your reference. And you can vote the idea here.

Regards