Python Eikon API - Insufficient Data

Hello,

I also see this issue when inside the codebook.

WARNING:pyeikon:Error with FXDKWEUR17=NPX: Insufficient data
ERROR:pyeikon:FXDKWEUR17=NPX: Insufficient data |

The weird thing is that there is actually no error, i can see the data inside the date frame. Any idea why it is coming up?

Find more posts tagged with

eikon-data-api

Accepted answers

zoya faberov

Hello @Moody ,

I have run your code several times, in CodeBook, I have reproduced the issue twice, for different instruments.

I suspect that the issue is not a specific content issue, as was identified and discussed in the previous discussion thread, but with the tight loop hitting 5 requests per second limit, from time to time, please see Eikon Data API Usage and Limits Guideline for the complete details.

I have introduced a brief pause after every 4 requests:

import datetime as dt
import pytz
import time

npast = 20    
sdate = (pd.datetime.now()+pd.DateOffset(days=-npast, normalize=True)).strftime("%Y-%m-%d")

allHist = pd.DataFrame()

ricbase = "FXDKWEUR"
ricend = "=NPX"

for i in range(1,25):
    ric = ricbase + str(i).zfill(2) + ricend
    print(ric)
    try:
        data = ek.get_timeseries([ric], start_date= sdate, fields="CLOSE").reset_index()
    except:
        print(f"failed for {ric}")
    hour = i-1 #str(i -1)
    dataF = data.rename({"CLOSE": hour}, axis=1)
    dataF_m = dataF.melt(id_vars = ["Date"], var_name = "hour")    
    allHist = pd.concat([allHist, dataF_m])
    if (i % 4) == 0:
        print('Sleep 2')
        time.sleep(2)

On my side, I have tested a couple of times and I am not hitting the issue anymore.

Please test- let us know how this works for you?

All comments

zoya faberov

Hello @Moody ,

Thank you.

To verify, inside CodeBook I run:

ek.get_timeseries(['FXDKWEUR17=NPX'], start_date = '2020-04-24', end_date = '2020-09-11')

My result was:

I was not able to reproduce this way -

Is this the exact same request that you run inside CodeBook? Is the error consistent on your side?

Next, with regard to this related discussion thread, as recommended by @pierre.faurel, if the same issue is reproducible in Eikon Excel, it will need to be approached and addressed, as content issue.

Moody

import datetime as dt
import pytz


npast = 20    
sdate = (pd.datetime.now()+pd.DateOffset(days=-npast, normalize=True)).strftime("%Y-%m-%d")


allHist = pd.DataFrame()

ricbase = "FXDKWEUR"
ricend = "=NPX"

for i in range(1,25):
    ric = ricbase + str(i).zfill(2) + ricend
    print(ric)
    try:
        data = ek.get_timeseries([ric], start_date= sdate, fields="CLOSE").reset_index()
    except:
        print(f"failed for {ric}")
    hour = i-1 #str(i -1)
    dataF = data.rename({"CLOSE": hour}, axis=1)
    dataF_m = dataF.melt(id_vars = ["Date"], var_name = "hour")    
    allHist = pd.concat([allHist, dataF_m])

hello,

could you please test this script inside your codebook. You will see an error run here:

WARNING:pyeikon:Error with FXDKWEUR17=NPX: Insufficient data
ERROR:pyeikon:FXDKWEUR17=NPX: Insufficient data |

However the data is in the dateframe.

zoya faberov

Hello @Moody ,

I have run your code several times, in CodeBook, I have reproduced the issue twice, for different instruments.

I have introduced a brief pause after every 4 requests:

import datetime as dt
import pytz
import time

npast = 20    
sdate = (pd.datetime.now()+pd.DateOffset(days=-npast, normalize=True)).strftime("%Y-%m-%d")

allHist = pd.DataFrame()

ricbase = "FXDKWEUR"
ricend = "=NPX"

for i in range(1,25):
    ric = ricbase + str(i).zfill(2) + ricend
    print(ric)
    try:
        data = ek.get_timeseries([ric], start_date= sdate, fields="CLOSE").reset_index()
    except:
        print(f"failed for {ric}")
    hour = i-1 #str(i -1)
    dataF = data.rename({"CLOSE": hour}, axis=1)
    dataF_m = dataF.melt(id_vars = ["Date"], var_name = "hour")    
    allHist = pd.concat([allHist, dataF_m])
    if (i % 4) == 0:
        print('Sleep 2')
        time.sleep(2)

On my side, I have tested a couple of times and I am not hitting the issue anymore.

Please test- let us know how this works for you?

Moody

thank you - that works !

EXPLORE OUR SITES