@Hubert CANEVET, the only reason why our sample codes delete objects immediately after results are received is to avoid cluttering up your environment with objects created by demo code. That does not mean you should proceed this way, your workflow should serve your own business and technical requirements.

Personally I'd use On Demand extractions for all asynchronous requests, instead of Scheduled extractions (if you are not familiar with the difference between these 2 methods, see here). When you use on demand extractions nothing is left lying around on the server, so there is no need to purge anything.

But if you want to use scheduled extractions and keep objects on the server for a certain time then you will have to create a purging mechanism. In that case I'd recommend you create a separate process to run in the background and purge old items. Usually garbage collection is not under stringent time constraints.

Hello,

Thank you for your answer.

I am afraid we are too advanced in the development to consider choosing another architecture, I remember I asked the question about calling certain report types synchronously, the answer did not encourage a lot to go that way.

So I am going to pursue using loops for the purging. Do you have an idea whether it is possible to launch several such treatments concurrently with the same account, in case one treatment is not quick enough ?

To launch queries it is possible with the API, I presume to delete the objects it should be the same.

@Hubert CANEVET, yes you should be able to delete objects using several parallel queries.

Oh, still one thing to verify : if I delete an extraction, that also deletes the files that come with it, does not it ?