get_timeseries throws if cannot get info for 1 ticker

igorg

{'fields': 'VOLUME;OPEN;CLOSE;HIGH;LOW', 'startDate': '2018-11-30T00:00:00', 'endDate': '2018-11-30T00:00:01', 'interval': 'daily', 'instruments': ['1ENYM9', '1ENYU9', '1ENYZ8', '1ENYZ9']} 

The function throws with the exception:

"Empty data passed with indices specified."

This comes from panda. It seems a check for empty data is missing in eikon module.

The problem is only 1 symbol (1ENYZ9) from the instrument list has no data. So if I pass 200 instruments to the function, I receive 0 results.

This looks like a bug. Could you confirm this?

Thanks,

Igor

Jirapongse

@igorg

Yes, it throws an exception. It is a bug because If I set raw_output to True, it properly returns valid JSON data.

df = ek.get_timeseries(rics=['1ENYM9', '1ENYU9', '1ENYZ8', '1ENYZ9'], 
                       start_date='2018-11-30T00:00:00', 
                       end_date='2018-11-30T00:00:01', 
                       interval="daily", 
                       raw_output=True)
df

However, the Python library is unable to parse this data.

The workaround is setting raw_output to True to retrieve the raw JSON data instead of DataFrame.

igorg

Thanks for your answer. However, I cannot accept this as a good workaround. As discussed here https://community.developers.refinitiv.com/questions/34580/json-strucutre-documentation.html with @Alex Putkov. , there is no public documentation available for raw_format. I would love to use it instead of the DataFrame, because it brings only the problems. But since that format is still under development and can be changed at any time, how can I use it in our production environment?

Will this bug be fixed in nearest future?

Zhenya Kovalyov

@igorg, I have reported this both to the product manager and to the dev team, and will let you know once I hear back from them.

igorg

Thank you. I will wait for some response from you.

For now, without any modifications, I can run only 1 instrument per request.. which is much slower, than 100 or 200 at a time.

Zhenya Kovalyov

@igorg since it throws when there have not been any trades for an exchange traded asset, here is a simple check you can do as a workaround:

df, e = ek.get_data(['1ENYM9', '1ENYU9', '1ENYZ8', '1ENYZ9'], ['HSTCLSDATE']) df[df['HSTCLSDATE'].notnull()]['Instrument'].tolist() 

What it does is looks at the historical close date. If the field is null, there have not been any trades:

igorg

what is HSTCLSDATE ? I don't see this field in Eikon DIB.

Zhenya Kovalyov

This field will have a value if there were trades any time before today. It gets populated during the 'closing run', when the value in the last trade price field gets copied to HST_CLOSE (or CF_CLOSE) and last trade price gets cleared.

Zhenya Kovalyov

That approach, of course, is valid if you are after inter-day time series. If you are looking for the intraday data, you can check whether there is a trade price or that the volume/turnover is not 0.

dshi

Any updates on this?