IMF: not comprehensive
As a user reported, the number of datasets available for the IMF on DBnomics is well below the numbers of datasets supplied by the API.
If we look at the Dataflow from the IMF API here http://dataservices.imf.org/REST/SDMX_JSON.svc/Dataflow, we could get all the datasets whose @id
doesn't contain a date as a principle. For example, we shouldn't take DS-FM201410
but we should take DS-FM
. Another example : we shouldn't take ̀DS-BOP_2017M06
but we should take DS-BOP
. As the format of the date in the code of the dataset is not always the same, we should just remove datasets whose code contains "20".