Explain the Error "Error code -1 | UDF Core request failed. Gateway Time-out"

gjastia
gjastia LSEG
edited June 4 in Refinitiv Data Platform Libraries

Hi Team, can you please explain the Error that the client is facing. They are pulling data for a list of ISINs and they are getting that error. Previously, for their bulk number of ISINs, when they remove the data item ['TR.NIIssuePricePctPrint','TR.NIIssuerUltParentTRBCEcoSec','TR.NIOfferPricePrintUniform'], the error is gone and the script works. However upon further testing, we can use the the 3 fields above for less number of ISINs.

Is this a limitation error since we did not find any documentation for that?

import refinitiv.data as rd
rd.open_session()
df = rd.get_data(
    universe = ['XS2624942764', 'US09710H5182', 'XS2591444422'],
    fields = [
#'TR.FiTicker',
'TR.LegalEntityIdentifier',
'TR.FiIssueDate',
'TR.FiCurrency',
#'TICKER',
#'TR.ExchangeTicker',
'TR.FiOriginalAmountIssued',
'TR.FiOrgID',

#'TR.OrgidCode',
'TR.FiParentOrgID',
#'TR.NACEClassification',
'TR.FiESGBondType',
#'TR.GreenBondFlag'#,
'TR.FiMaturityStandardYield',
'TR.NIIssuePricePctPrint',
'TR.NIIssuerUltParentTRBCEcoSec',
'TR.NIOfferPricePrintUniform']

)

display(df)

image-64bf746d8f51e-d140.jpeg

 

Tagged:

Answers

  • jason.ramchandani01
    jason.ramchandani01 admin

    Hi @gjastia so that message usually means that the request was not completed by the server. Reasons for this could be size of request (requests to the service should not be more that 10,000 datapoints per API request - please see our limits guidance here). I can see that their request is already over that limit with 1756 rows x 12 columns = 21072 - almost twice the recommended amount of datapoints. When server load is light these queries might pass but during busier periods they may well fail - hence our guidance around the 10K datapoint per call. The request also seems to be mixing reference fields with timeseries fields. There is also no row fidelity for the timeseries fields as they are using the get_data function - they should use the get_history function to ensure row fidelity across different timeseries. They can open the Codebook App and look in the Examples folder there and check the get_history function as well as get_data function examples etc. I hope this can help.

Categories