bug in get_timeseries?

tsando · May 2016

Hi,

If I run the following command in python:

dic = ek.get_timeseries(["CFI2Zc1",  # Carbon
                         "LCOc1"
                         ],
                        start_date=datetime.datetime(2008, 1, 1),
                        end_date=datetime.datetime(2016, 12, 31),
                        interval='daily',
                        fields=['CLOSE'])

I don't get a dataframe that starts from 2008, but 2010 instead. The more RIC codes I add, the less the history goes back. Is this behaviour normal? I only seem to get the correct dates when only using a single RIC code.

Denis Dollfus · May 2016

Hi @tsando , I confirm the problem. The maximum number of points for inter-day data is currently 3000, unfortunately to be divided by number of instruments. So with 10 instrument the maximum becomes 300 points.

This behavior being error-prone, I'll raise this issue to time series service.

Thanks for your feedback.

iain.scott · May 2016

Bear in mind that this service is EAP with the main aim of validating the APIs . Given the service is not production ready there may be some issues with underlying data that we would expect to resolve before launch.

Thank you for the feedback, we will look into the issue.

vlscout · May 2017

1) I confirm that this restriction still exists.

2) Also, the timeseries function returns an error (AVTCHB.ST', ': ', 'Error: TSIUnknownInstrument, ErrorCode: TA-TSIUnknownInstrument, Fault: TSIError, Description: Invalid RIC', '\n')

for RICs that contain lower case letters like AVTCHb.ST, ANWGn.F etc.

3) And for recent IPOs that have a limited trading history, if you request a time series outside of that history, the function will return that recent trading data regardless whether it matches your requested time frame.