On Mon, 11 Feb 2019, klaus.hasenbach(a)web.de wrote:
Thank you. You are right
Problem 2.) I did not got the error message again.
Problem 1.) So slow that I get no anser at all when I try to retieve
> 1000 time series at once what I did before easy.
I will wait and try to retrieve the data some days later and hope that
the server of dbnomics will run better then.
Nevertheless, is the a chance to run gretl with the old extension of
dbnomics? Just to test the possibilty that the new version might be also
a problem. The version I run now is from 2019-01-17
Interesting question. It seems the dbnomics v21 API is still
supported for now (though we don't know for how long). I've put a
compatible version of dbnomics.zip at
http://ricardo.ecn.wfu.edu/pub/gretl/dbnomics_v21/dbnomics.zip
and you can try it out if you wish.
One thing I've noticed: under the v21 API if you asked for actual
data you mostly got actual data, with only the necessary metadata.
Under the current v22 API you get a ton of metadata, whether you
want it or not. I can see that slowing downloads, and it also means
that our JSON parser has a lot more work to do in extracting what's
wanted from the copious stream received.
There was a comment on this on the dbnomics forum
https://forum.db.nomics.world/t/dbnomics-api-v22-released/204/4
but nobody has made an issue of it. Perhaps we should raise the
subject again.
Allin