Why do I get only N/A in the following code in Python?

susskaya.anita

import refinitiv.dataplatform.eikon as ek

import numpy as np # NumPy

import pandas as pd # pandas

import configparser as cp

import sys

ek.set_app_key('') # here I insert app_key

import eikon as ek

df_xls = pd.read_excel(r'C:\10K.xlsx')

mylist = df_xls['RIC'].tolist()

#print(mylist)

df, err = ek.get_data(

instruments = mylist,

fields = ['MEDIAN(TR.PtoEPSMeanEst(Period=NTM,Frq=W,SDate=2016-12-15,EDate=2021-12-14))']

)

display(df)

Below you could see the result of the script:

umer.nalla

Hi @susskaya.anita , @malashenko.i.o , @gabriel.g and @rombout

I have been advised by the Eikon Data API team that this issue has been fixed and passed their internal testing with 3000 RICS.

Please test and confirm. Thank you for your patience.

raksina.samasiri

Hi @susskaya.anita

I used the code you provided and can get the data properly

Could you please try to print an error of get_data function using print(err) ?

This could be the case that you don't have an access to this data, if so, please contact your account owner regarding the entitlement or if it's any other issue, you may put the error message here and we will investigate more.

susskaya.anita

[{'code': 100, 'col': 1, 'message': 'Asynchronous Query library internal error.', 'row': 0}, {'code': 100, 'col': 1, 'message': 'Asynchronous Query library internal error.', 'row': 1},...

it is the error

susskaya.anita

It seems that it does not uderstand properly the values from xls. I don't know why.

susskaya.anita

I don't think that the problem is in the access because via Excel add-in I could retrieve that data

umer.nalla

Hi @susskaya.anita

Your XLS filename suggests you have 10K instruments in the list. If this is the case, have you tried with a smaller list of instruments in an excel file?

If you have tried a smaller list and it still fails, please convert the shorter excel file to a CSV file and attach it here so the above code snippet can be tested with your list.

susskaya.anita

You were right. The decrease of number of RIC-codes helped. However, what is the limit? Only 100 or what?

umer.nalla

Hi @susskaya.anita

The following page explains the various types of limits enforced with the Eikon Data API

Eikon Data API Usage and Limits Guideline

susskaya.anita

get_data: The current limit value (10-Oct-2019) is around 10,000 data points.

But I did not exceed that limit. Moreover, I did not have errors which are mentioned in the guideline

malashenko.i.o

raksina.samasiri, I have the same issue starting this week and I had some research. This issue reproduces time from time and possibility depends on the amount of data.

The code:

for RIC_num in [10,100,500]:

for n_cols in [10, 20]:

RIC_list = ric.RIC.tolist()[:RIC_num]

columns = columns_list[:n_cols]

print('RIC length:', len(RIC_list))

print('columns length:', len(columns))

for i_ in range(10):

data, err = ek.get_data(RIC_list, columns)

if err is None:

print('No error', datetime.now())

else:

print(set([e['code'] for e in err]), end=' ')

if 100 in set([e['code'] for e in err]):

print('<--- This error!', datetime.now())

else:

print(datetime.now())

The output (partly):

RIC length: 100
columns length: 20
{251658243} 2021-12-16 17:47:00.936565
{251658243} 2021-12-16 17:47:08.035268
{100} <--- This error! 2021-12-16 17:47:30.244498
{100} <--- This error! 2021-12-16 17:47:50.696324
{251658243} 2021-12-16 17:47:57.530822
{251658243} 2021-12-16 17:48:05.015905
{100} <--- This error! 2021-12-16 17:48:40.422272
{251658243} 2021-12-16 17:48:50.582018
{251658243} 2021-12-16 17:48:58.884504
{251658243} 2021-12-16 17:49:08.098476

Here you can see 3 fails for 100 RICs and 20 columns.
In case of 500 RICs with 20 columns it fails in 8 out of 10 attempts:

RIC length: 500
columns length: 20
{100} <--- This error! 2021-12-16 17:50:49.137280
{100} <--- This error! 2021-12-16 17:51:10.376004
{251658243} 2021-12-16 17:51:32.593595
{251658243} 2021-12-16 17:51:55.249686
{100} <--- This error! 2021-12-16 17:52:16.634033
{100} <--- This error! 2021-12-16 17:52:37.690012
{100} <--- This error! 2021-12-16 17:52:58.664758
{100} <--- This error! 2021-12-16 17:53:19.581651
{100} <--- This error! 2021-12-16 17:53:40.567594
{251658243} 2021-12-16 17:54:03.253674

I used to work with 3000 rows and it worked fine till before.

umer.nalla

Hi @susskaya.anita

I have reported this issue to the Eikon Data API team for their input.

I will report back once I hear from them.

susskaya.anita

Thank you!

umer.nalla

Hi @susskaya.anita and @malashenko.i.o

The EDAPI team have asked if you can generate logs?

import logging
ek.set_log_level(logging.DEBUG)
ek.set_app_key('your app key')

and then running your script should generate a file something like pyeikon.20211217.15-28-56.log

please remove any credential information from the log file - before uploading.

e.g.

headers = {'x-tr-applicationid': 'your appkey'}

and

access_token":"eyJhbGciOiJIUzI1NiIsInR.......'

NOTE: There may well be multiple entries like the above in your log file

please search and remove all of them before uploading.

malashenko.i.o

@umer.nalla pyeikon.20211217.18-58-35.log.zip

umer.nalla

Hi @malashenko.i.o - thanks will pass this on.

gabriel.g

I am experiencing similar issues.

Snippets that worked without any problems for years keep throwing this error when calling ek.get_data:

Error code 100 | Asynchronous Query library internal error.

As far as I can tell the problems started on December 10th.

Lowering the number of instruments/fields seems to reduce the error rate. However, even when using moderately sized requests the mentioned error occurs occasionally.

susskaya.anita

the same error I usually see

susskaya.anita

pyeikon.20211217.18-58-35.zip

umer.nalla

Hi @susskaya.anita , @malashenko.i.o and @gabriel.g

I have copied your messages and log files to the Eikon API team.

I will update once I hear from them.

susskaya.anita

Thank you! I will wait then!

susskaya.anita

Thank you so much for raising a ticket for me!

lukasz.trzcinski0

I checked the script using 1500 RICs and get NA as well. Tried the same with TR.PtoEPSMeanEst only and get all results.

My assumption is that for MEDIAN(TR.PtoEPSMeanEst(Period=NTM,Frq=W,SDate=2016-12-15,EDate=2021-12-14)) calculation, python needs to get all data points to calculate MEDIAN, so 261 separate values for each RIC that is why the limit is exceeded. Why the error message doesn't indicate the exceeded limit? I don't know,...

malashenko.i.o

I don't think there is a limit restriction, because, as I posted below, you could try to get the same data in a loop and it could work sometimes (possibility depends on how big your query is). In case of limits there would be no chance for this.

susskaya.anita

As you know via Formula Builder I could get results using this formula for 7500 RIC-codes. I don' think that via Formula Builder we could get more than via Python.

lukasz.trzcinski0

So why if we change the Frq from W to Y we can get all the results? I tested it for 1500 RICs and worked well

umer.nalla

Hi @lukasz.trzcinski0 - Are you referring to two different limits - one being the Eikon API limits as defined on the https://developers.refinitiv.com/en/api-catalog/eikon/eikon-data-api/documentation and another being the limitation of the MEDIAN call - on the backend system that EDAPI uses perhaps?

@malashenko.i.o , @gabriel.g and @susskaya.anita - Can you please confirm which country you are located in - so the EDAPI team can confirm which servers your EIKON is connecting to. - in case there is a server-side issue.

UPDATE: No need for pip freeze output. It appears the problem is server-side.

Please still confirm your location so we can identify the servers.

Thanks.

susskaya.anita

Russia, Saint Petersburg

malashenko.i.o

@umer.nalla Russia

gabriel.g

Austria, Vienna
connecting via EMEA servers

umer.nalla

Hi @susskaya.anita , @gabriel.g and @malashenko.i.o

Your tickets should be re-assigned to the server team - who will work with the EDAPI team to investigate further.

Hopefully, the support team will keep you updated on the progress.,

rombout

Has this already been resolved? I am getting similar error messages:

"Asynchronous Query library internal error."