Is there an interval type 30 minutes for python api?

nils.engelhardt

Is there an interval type 30 minutes for python api?

Alex Putkov.1

To create timeseries in 30 minute interval you'd need to retrieve 1 minute bars and summarize them into 30 minute bars.

pierre.faurel

Hi,

If you wonder if get_timeseries function accepts other interval type than 'tick', 'daily', ... , the response is no.

nils.engelhardt

thanks a lot!

wyj

Thanks ,after i get data via

df1=ek.get_timeseries(s1,fields=["Open","High","Low","Close","Volume"], start_date = "2017-11-13T01:00:00",end_date = "2017-11-13T07:04:05",interval='minute')

Do u have any high effect ways to handle this data into 5mins bar?

wyj

I hope there got 5mins, 30mins,hour.....

Alex Putkov.1

There's nothing from Thomson Reuters that would facilitate this operation, since Python has very rich capabilities for all kinds of data transformation, summarization and statistical analysis.

pierre.faurel

This a simple example to extract a 30 minute interval timeseries from a minute interval result:

import eikon as ek
from datetime import timedelta

interval = 30

def ohlcsum(df):
    return [df['OPEN'][0],df['HIGH'].max(),df['LOW'].min(),
           df['CLOSE'][-1],int(df['VOLUME'].sum())]

data_1min = ek.get_timeseries('PEUP.PA', interval='minute')
timestamps=[]
result=[]
start_slice = data_1min.index[0]
while True:
    end_slice = start_slice + timedelta(minutes=interval-1)
    interval_slice = data_1min.loc[start_slice : end_slice]
    try:
        data = ohlcsum(interval_slice)
    except IndexError:
        break
    timestamps.append(interval_slice.index[-1])
    result.append(data)
    start_slice = start_slice + timedelta(minutes=interval)
    if start_slice > data_1min.index[-1]:
        break

labels = ['OPEN','HIGH','LOW','CLOSE','VOLUME']
data_new_interval = pd.DataFrame.from_records(result, index=timestamps, columns=labels)

wyj

I don't think u got the right data for 5 min

for every 5 min data, it should be calculated by the following function

def ohlcsum(df):
    df = df.sort()
    return {
       'Open': df['Open'][0],
       'High': df['High'].max(),
       'Low': df['Low'].min(),
       'Close': df['Close'][-1],
       'Volume': df['Volume'].sum()
      }

wyj

Alex, thanks for your answer. Of course You are right since Pyhon has very rich libs for calculation, but my requirement is to get 3000 RICs' 5mins bar in real-time, do you have any efficient way to handle this? thanks again

Alex Putkov.1

Well, your use case is borderline outside what a market data terminal can handle. If indeed you're looking to construct 5 min bars in real-time, i.e. form the continuation of the bars live from the real-time data stream, then you definitely cannot do this using hosted (aka Internet delivered) Eikon, as your concurrent real-time data subscriptions are limited to 2.5K RICs.

Alex Putkov.1

Eikon Python API currently does not provide streaming market data at all. But even if you used Eikon .NET API, which does provide 5 min bars and live continuation from real-time data stream, you'd still be subject to 2.5K RICs cap on the number of concurrent real-time data subscriptions.

Alex Putkov.1

Using Eikon your only option would be to poll the timeseries service every 5 minutes for a new 5 min bar if you use .NET API or for 5 new one minute bars if you use Python API. This won't give you the updated 5 min bar in real-time, there certainly will be some delay. How well this will work for 3K RICs if you do it continuously every 5 minutes? There's only one way to find out, but my guess is probably not fully reliably. You should be prepared that some requests will fail and you might need to resubmit, which would result in further delay in obtaining the new bar.