Optimized Way to download specific files from Azure blob Storage

[jaryszek]

hi Guys,

i have structure in Azure Blob Storage like
container1
-folder1
-file1
-file2
-folder2
-file3

And now what is the best optimized way to get data using power query?

I need to connect to different folder files in each seperated queries.

How to optimize it?

so for example i need one table from folder2 which is file3. It is one query

but also i need file1 and file2 in 2 other queries.

How to make from this best performance? How to filter table ? First get all azure blob storage data and filter it by each query?

Thank you for your advice.

[v-kpoloju-msft]

Hi @jaryszek,
Thank you for reaching out to the Microsoft fabric community forum. 

To achieve the best performance when connecting to specific files in Azure Blob Storage via Power Query, it's important to avoid loading the entire container or folder and then filtering afterward. Instead, follow this optimized approach:

Connect to Azure Blob Storage from Power BI using: Home → Get Data → Azure → Azure Blob Storage. After connecting and authenticating, select your container (e.g., container1) and click Transform Data.

In the Power Query editor, you’ll see metadata including Name, Folder Path, Extension, and Content. Apply early filtering (before expanding or transforming), using the Folder Path and Name columns to load only the desired file.
Here's an example for loading only file1 from folder1:

let

    Source = AzureStorage.Contents("https://yourstorageaccount.blob.core.windows.net"),

    Container = Source{[Name="container1"]}[Content],

    Filtered = Table.SelectRows(Container, each [Folder Path] = "https://yourstorageaccount.blob.core.windows.net/container1/folder1/"and [Name] = "file1.csv"),

    FileBinary = Filtered{0}[Content],

    Data = Csv.Document(FileBinary)

in

    Data

This way, each query loads only the specific file it needs, resulting in much better performance and reduced load times.

If you're still facing issues or the structure is more dynamic, feel free to share more details like file types or folder structure we’ll be happy to help you fine tune the query.

Thank you for using the Microsoft Community Forum.

[jaryszek]

Ok i put 50 gigs (100 different files)  to one container. 

I went to Get data --> Azure Blob Storage --> Load. 

And imagina that it is reading all the files immidietely!! It sucks. 

So it means that it is loading 50 GB of data eventhough I am not importing them into seperated queries!!:

How to make this working fast and reliable? 
Best,
Jacek

[jaryszek]

"Sure! The most optimized way is to connect directly to each file you need by using its exact path instead of loading the full container and filtering inside Power Query. That way, Power Query only fetches what's needed, making it faster and lighter. I can help you set it up properly if you want."

"No worries! You’re right—AzureStorage.BlobContents needs a proper path and context. I can help you get the exact file using the right method without relying on Web.Contents. Just send me the folder structure and file names, and I’ll guide you step by step."

this is what i got from scamer on facebook. But it really possible to connect to only one file using azureblob.contents method? 

Best,
Jacek

[lbendlin]

I don't think you will see any wiggle room for performance enhancements there.  Can you show a sample Power Query code?  How large are these blobs?

[v-kpoloju-msft]

Hi @jaryszek,

Thank you for reaching out to the Microsoft fabric community forum. 

As mentioned by the community member @lbendlin, could you please provide a sample Power BI query? This will help me better understand your issue and provide an appropriate solution.

Thank you.

[jaryszek]

@v-kpoloju-msft  @lbendlin  there i no query yet. I am designing architecture for power query. 

How to make the implementation? Assuming that blob can be huge in the future. 
There are just csv files which i need to get into power query but i didn't find a way to only download one file from blob storage -> it has to be always filtered with all files first. But this is a good approach? 

Or maybe just download every file and after that build queries on top of that using filter to direct file? 

Thank you for your help,
Jacek