Get time series, start dates don't match.

Options

Best Answer

  • jason.ramchandani01
    jason.ramchandani01 admin
    Answer ✓

    @pyquant Here is the generic routine:

    from dateutil import parser
    from datetime import timedelta
    from datetime import datetime
    import math
    import time
    def date_range(start, end, intv):
        start = datetime.strptime(start,"%Y-%m-%d")
        end = datetime.strptime(end,"%Y-%m-%d")
        diff = (end  - start ) / intv
        for i in range(intv):
            yield (start + diff * i).strftime("%Y-%m-%d")
        yield end.strftime("%Y-%m-%d")
    def get_daily(rics,fields,start,end):
        for ric in rics:        
            interval = math.ceil((parser.parse(end) - parser.parse(start)).days / 3000)
            l = list(date_range(start,end,interval))
            df1 = pd.DataFrame()
            df = pd.DataFrame()
            for i in range(interval):
                ts = ek.get_timeseries(rics=ric,fields=fields, start_date=l[0+i],end_date=l[1+i], interval='daily')
                df = df.append(ts)
                time.sleep(0.4)

        return df
    rics = ['.GDAXI'] # Just for one ric at the moment I will extend this for multi-ric
    fields = ['OPEN', 'HIGH', 'LOW', 'CLOSE']
    start = '1990-06-04'
    end = '2018-06-04'
    df = get_daily(rics,fields,start,end)
    df

    I hope this can help.

Answers

  • jason.ramchandani01
    jason.ramchandani01 admin

    Hi @pyquant Thanks for your question - the reason is because the timeseries API is limited to 3000 rows per interday request. You can see the limits for each in this document.

    You can of course iterate these calls as you are already doing for each RIC. But you can also achieve this with 2 calls per RIC.

    df1 = ek.get_timeseries(i, ['CLOSE'],start_date="1999-01-02", end_date="2008-10-22")
    df2 = ek.get_timeseries(i, ['CLOSE'],start_date="2008-10-23")
    df = pd.concat([df1, df2], axis=0)

    image

    I hope this can help - I have a generic routine for the date calcs somewhere I will add it when i find it.


  • pyquant
    pyquant Newcomer

    Thanks. I will give it a try.

Categories