Refresh cache for API direct import

[damanpradhan]

Hi,

So I have a scenario where I have to use API's. 

Right now, I am accessing the API's directly(as they are within the same domain), but in a few days, there will be an API gateway that will provide access to these APIs. 

My requirement is that I should be able to:

1. Access the APIs and build the dataset

2. Have some way to periodically refresh the dataset from the APIs, ideally adding to the already existing data.

I have searched extensively for any literature on how PowerBi deals with any from of cache for API direct imports. However, I haven't come across any.

Can anyone help me on this? 

[v-juanli-msft]

Hi @damanpradhan 

I'm not clear about these:

"there will be an API gateway that will provide access to these APIs"

"cache for API direct imports"

If you connect to a web api with Power BI Desktop, you could configure schedule refresh in this way

https://sqlserverbi.blog/2018/10/21/web-api-data-sources-with-power-query-and-scheduling-data-refres...

If your api is an secure api which needs authentication,

Power BI – Connect to your secure API

How to refresh ODATA API with OAuth2

Best Regards
Maggie

Community Support Team _ Maggie Li
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[damanpradhan]

Hello @v-juanli-msft,

Thank you for your response. However, what I am looking for is a little more in-depth guidance. 

Allow me to break that down for you:

 I have to access data via APIs. The data is on the larger scale and since it is more of a direct import for PowerBi, I do not want the end user experiencing delays in PowerBi because he chose a large time scale from a time slicer and the data is taking time to come through the wire to PowerBi.

Question: When we import data via APIs, does PBI make use of some storage that we can refresh periodically? That way I can bring in the whole data upfront to the cache and make the PowerBI experience for the user a satisfactory one.

Question: Does PBI have some heuristics that determine that if an API has been previously called, then it can reach out to a specific cache that will contain the same data?

Question: If the answer to the above question is yes, then does it mean that there will be multiple caches the different API calls? If yes, then how can we see the details and preferably have more control on each?

Hope you can help me answer these questions and also suggest alternatives to solving this kind of a situation where the data ingress is very large.

Thanks in advance ! 

(@v-juanli-msft  Tagging you)