[MainThread 47708] Backend error. 400 Bad Request

mcooper

I am attempting to download some data using the Eikon python package (version == 1.1.2 due to some issues with more recent versions), however it is very hit or miss whether I actually receive a response from the Eikon servers. Very frequently I get the 400 error (bad request) and when I do receive a response frequently many, if not all, the values in my resulting dataframe are "na". Is there anything I can do to ensure I receive prompt and complete data?

Code to Reproduce:

import eikon as ek

ek.set_app_key("YOUR KEY HERE")

for years_back in range(25):
    year = str(2020 - years_back)
    print(year)
    s, e = 365 * years_back, 365 * years_back + 1
    data, err = ek.get_data(tickers,
                            fields=['AVG(TR.PriceToSalesPerShare(SDate=-{s},EDate=-{e},Frq=D))'.format(s=s, e=e),
                                    'AVG(TR.CompanyMarketCap(Curn=USD,SDate=-{s},EDate=-{e},Frq=D))'.format(s=s, e=e),
                                    'AVG(TR.DailyValueTraded(Curn=USD, SDate=-{s},EDate=-{e},Frq=D))'.format(s=s, e=e),
                                    'TR.CoRPrimaryCountry'])

Environment:

conda version : 4.8.4

pip: 20.2.2

platform : win-64

eikon: 1.1.2

pandas: 1.1.1

Alex Putkov.1

This error is most likely due to timeouts on the backend. The requests you're submitting are fairly taxing on the backend, which is required to retrieve daily timeseries for each of the fields and aggregate them (calculate the average). I would suggest to try reducing the number of tickers in each request. I appreciate this may increase the overall retrieval time, but this is pretty much the only thing you can do from the client side to improve the success rate of these requests. If this workaround is not acceptable, then I would suggest using Fiddler to capture a few requests that return "Bad request" or nulls, and then opening a support case with Refinitiv Helpdesk, so that Refinitiv Support could investigate the backend errors.