Cdsapi request without download


I want to use the cds API to make a data request, but I don't want to download the data until much later. Is it possible to separate the request and the download parts of a cds api "retrieve" call?

Currently I send the retrieve call, check the website to make sure the request is in the queue, but then have to crash out of the python script with ctrl-c. Otherwise the process just sits and waits for hours while the request is being filled.

Ideally I'd like to place a request without downloading. At some later time, I'd then place a new retrieve call to do the actual downloading.

Any help or direction is appreciated.

Hi Tim,

FYI The data you request are copied into the CDS cache before they are downloaded. As they may only be kept in the cache a short time  (depending on the load on the CDS) before being automatically deleted to recover space, it may be better to leave the script running so that the data are downloaded as soon as they are available.

Hope that helps,


Hi Kevin,

I understand the cache usually holds data for at least a day and a half. My requests are taking up to 12 hours. I'd rather not tie up a processor while it waits for that 12 hours, but I'm happy to poll the service every 6 hours or so to see if the data are ready. The RDA archive separates requests from downloads, and I was hoping I could do the same with CDS.