Reading parquet files from blob storage

[Anonymous]

Hi,

I have a service on Azure working called Time Series Insights. This service stores data into a blob storage in a .parquet format. I would like to access this data from Power Bi.

When I connect to the blob storage however I am only given 'meta data' on what is in the container, not the actual data in the .parquet. See below:

What can I do differently, or what haven't I done, to access and create visualisations of the data within the .parquet?

Any advice on how to move forward would be appreciated.

Thanks

[Greg_Deckler]

The flavor says "butter"

[amitchandak]

Refer, if this can help

http://www.johndehavilland.com/blog/2019/06/10/parquet-files-with-powerbi.html

https://community.powerbi.com/t5/Desktop/Reading-Data-in-PowerBI-from-parquet-format-file-from-ADLS-...

[Anonymous]

Unfortunately the first link's solution using Apache isn't scalable in my application. I need the PBI dashboard to do it natively or using cloud resources.

The second link deals with ADLS instead of blob storage. Unfortunately Time Series Insights only sits on a blob storage component so I can't switch to ADLS Gen2. 

Any way for me to achieve this? 

[Greg_Deckler]

@Anonymous , more seriously though, .parquet files are often associated with rubbish Python code so I would recommend maybe using the pandas library in a Python query step perhaps:

https://arrow.apache.org/docs/python/parquet.html

[Anonymous]

@Greg_Deckler I have no control over the file format. Files stored on the account for Time Series Insights are sent directly from eventhubs. Files start as .json and get sent through Azure functions through eventhubs into Time Series Insights storage account.

[v-kelly-msft · Accepted solution]

Hi @Anonymous ,

1. click Transform Data:

2.Then,click on Binary just to double check your data.

3.Click on To Table. A window will open, click Ok.

4.Click on the arrows to the right of the column named “Column1”. Here, uncheck the optionUse original column name as prefix - this will add unnecessary prefixes to your variable names. If you have an alert with List may be incomplete, click on Load more… and then Ok.

5.Your dataset will finally appear. Click on Close and Apply, and wait for the changes to apply. Your data is now ready for any Power BI dashboard!

Here is the reference.

Best Regards,
Kelly

Did I answer your question? Mark my post as a solution!