question

Upvotes
Accepted
1 0 0 1

Determine the ideal number of fields and assets to download in each query, to avoid null data or empty dataframes

The purpose of this case is to determine the ideal number of fields and assets to download in each query, to avoid null data or empty dataframes. I have noticed that when I try to download various metrics and assets, sometimes the data returned is null or the dataframes are empty. However, this issue is variable; I can run the same code again, and in the second iteration, the data might be complete. Additionally, the issue varies with RICs: for some RICs, this problem occurs more frequently than for others.

Moreover, I have observed that the same set of metrics may work fine for a given set of tickers, but if I add another ticker to that list, problems might arise. This inconsistency is also variable, as sometimes it occurs and sometimes it does not.

It is also likely that the problem is not in the code itself. The code returns a dictionary of dictionaries, where each key is a RIC and the values are dataframes containing the metrics. I compare each iteration to identify inconsistencies or issues with the data retrieval process. Do you know why is this happening? or if you know the most efficient way to retrieve huge lists of fields and RIC´s.

# Libraries
import pandas as pd
import time
import logging
import refinitiv.data as rd
rd.open_session()


#logging config
logging.basicConfig(level=logging.INFO, format='%(message)s')


# Logger Config
logger = logging.getLogger('myAppLogger')
logger.setLevel(logging.INFO)
formatter = logging.Formatter('%(message)s')


pd.set_option('future.no_silent_downcasting', True)


def rename_duplicate_columns(df: pd.DataFrame) -> pd.DataFrame:
    """
    Renames duplicate columns in the dataframe by
    appending '_loc_curr' and '_usd'
    to the first and second occurrence of each
    duplicate column name, respectively.


    Parameters
    ----------
    df : pd.DataFrame
        The dataframe containing columns that need
        to be renamed if duplicates are found.


    Returns
    -------
    df : pd.DataFrame
        The dataframe with renamed columns to
        handle duplicates.
    """
    cols = pd.Series(df.columns)
    for dup in cols[cols.duplicated()].unique():
        cols[cols[cols == dup].index.values.tolist()] = [dup + '_loc_curr', dup + '_usd']
    df.columns = cols
    return df




def is_data_valid(data: pd.DataFrame):
    # I check if the DataFrame obtained is valid (not empty and without only nan, since it can happen)
    return not data.empty and not data.isnull().all().all()




def retrieve_and_process_fundamental_data(tickers: list[str],
                                          metric_codes: list[str],
                                          ticker_partition_size: int,
                                          metric_partition_size: int,
                                          max_attempts: int,
                                          ):
    
    metrics_df = {ticker: pd.DataFrame() for ticker in tickers}
    fund_data_error_metrics = []
    fund_tickers_error = []
        
    for ticker_index in range(0, len(tickers), ticker_partition_size):
        ticker_partition = tickers[ticker_index:ticker_index + ticker_partition_size]
        logger.info(f"Retrieving data for tickers: {ticker_partition}")


        for metric_index in range(0, len(metric_codes), metric_partition_size):
            partitioned_metrics = metric_codes[metric_index:metric_index + metric_partition_size]
            logger.info(f"Retrieving data for tickers: {ticker_partition} and metrics: {partitioned_metrics}")
            time.sleep(0.2)
            logger.info("")
            
            try:                
                for attempt in range(max_attempts):
                    logging.info(f"Attempt {attempt + 1} to fetch data.")
                    metrics_data = rd.get_data(tickers, metric_codes)
                    if is_data_valid(metrics_data):
                        logging.info("Successful query.")
                        break
                    else:
                        logging.error(f"Attempt {attempt + 1} failed. Retrying...")


                for ticker in ticker_partition:
                    ticker_data = metrics_data[metrics_data['Instrument'] == ticker]
                    if metrics_df[ticker].empty:
                        metrics_df[ticker] = ticker_data
                    else:
                        metrics_df[ticker] = pd.concat([metrics_df[ticker], ticker_data], axis=1)
                    
                    # Format
                    metrics_df[ticker] = metrics_df[ticker].rename(columns={'Period End Date': 'Date'})
                                        
                    # Removing duplicates for specific columns
                    columns_to_check = ['Date', 'Instrument', 'Income Statement Orig Announce Date']
                    for col in columns_to_check:
                        if col in metrics_df[ticker].columns:
                            first_col = metrics_df[ticker][col].iloc[:, 0] if isinstance(metrics_df[ticker][col], pd.DataFrame) else metrics_df[ticker][col]
                            metrics_df[ticker] = metrics_df[ticker].drop(columns=[c for c in metrics_df[ticker].columns if c == col][1:])
                            metrics_df[ticker][col] = first_col
                    
                    metrics_df[ticker] = rename_duplicate_columns(metrics_df[ticker])
                    metrics_df[ticker] = metrics_df[ticker].infer_objects(copy=False)
                    logger.info(f"Done for {ticker}")
            except Exception as e:
                logger.error(f"Error retrieving data for tickers: {ticker_partition} and metrics: {partitioned_metrics} - {e}")
                fund_data_error_metrics.append(partitioned_metrics)
                fund_tickers_error.append(ticker_partition)
    
    return metrics_df




ticker_partition_size = 10
metric_partition_size = 15
max_attempts = 3


tickers = [
    "MSFT.OQ",
    "AAPL.OQ",
    "NVDA.OQ",
    "GOOGL.OQ",
    "AMZN.OQ",
    "META.OQ",
    "UNH.N",
    "BRKa.N",
    "LLY.N",
    "2330.TW",
    "AVGO.OQ",
    "NOVOb.CO",
    "V.N",
    "TSLA.OQ",
    "XOM.N",
    "WMT.N",
    "0700.HK",
    "MA.N",
    "CSCO.OQ",
    "PG.N",
    "005930.KS",
]


fields = [
    "TR.F.ComStockBuybackNet(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0)",
    "TR.F.NetIncAfterTax(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0)",
    "TR.F.ShrUsedToCalcDilEPSTot(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, ReportingState=Orig, Period=FQ0)",
    "TR.F.IncAvailToComShr(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0)",
    "TR.F.IncAvailToComShr(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0,Curn=USD)",
    "TR.F.DebtInclPrefEqMinIntrTot(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0)",
    "TR.F.CashSTInvstTot(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, ReportingState=Orig, Period=FQ0)",
    "TR.F.BookValuePerShr(SDate=2023-05-01,EDate=2024-05-30,Period=FQ0,Frq=FQ)",
    "TR.F.MinIntr(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, ReportingState=Orig, Period=FQ0)",
    "TR.F.DebtTot(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, ReportingState=Orig, Period=FQ0)",
    "TR.F.OpProfBefNonRecurIncExpn(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, ReportingState=Orig, Period=FQ0)",
    "TR.F.OpProfBefNonRecurIncExpn(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, ReportingState=Orig, Period=FQ0,Curn=USD)",
    "TR.F.TotAssets(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, ReportingState=Orig, Period=FQ0)",
    "TR.F.NetCashFlowOp(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, ReportingState=Orig, Period=FQ0)",
    "TR.F.DebtLTTot(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, ReportingState=Orig, Period=FQ0)",
    "TR.F.ShHoldEqCom(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, ReportingState=Orig, Period=FQ0)",
    "TR.F.ShHoldEqCom(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, ReportingState=Orig, Period=FQ0,Curn=USD)",
    "TR.F.EBITDA(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, ReportingState=Orig, Period=FQ0)",
    "TR.F.EBITDA(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, ReportingState=Orig, Period=FQ0,Curn=USD)",
    "TR.CashFromOperatingAct(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,Period=FQ0,Curn=USD)",
    "TR.CashFromOperatingAct(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,Period=FQ0)",
    "TR.NetIncomeBeforeExtraItems(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,Period=FQ0)",
    "TR.NetIncomeBeforeExtraItems(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,Period=FQ0,Curn=USD)",
    "TR.AssetTurnover(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,Period=FQ0)",
    "TR.LTDebtToTotalAssetsPct(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,Period=FQ0)",
    "TR.SaleIssuanceOfCommon(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,Period=FQ0)",
    "TR.SaleIssuanceOfCommon(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,Period=FQ0,Curn=USD)",
    "TR.GrossProfitMarginIndustrialAndUtilityPct(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,Period=FQ0)",
    "TR.F.TotCurrAssets(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0)",
    "TR.F.TotCurrLiab(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0)",
    "TR.F.LeveredFOCF(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0)",
    "TR.F.LeveredFOCF(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0,Curn=USD)",
    "TR.F.AvgNumShrOutst(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, ReportingState=Orig, Period=FQ0)",
    "TR.F.ShrUsedToCalcDilEPSTot(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, ReportingState=Orig, Period=FQ0)",
    "TR.ShortInterest(SDate=2014-01-01,EDate=2024-05-30,Frq=FQ)",
    "TR.F.TotRevenue(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0)",
    "TR.F.TotRevenue(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0,Curn=USD)",
    "TR.F.IntrExpnNetOfIntrInc(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0)",
    "TR.F.NetDebt(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0)",
    "TR.F.ComShrOutsTot(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, ReportingState=Orig, Period=FQ0)",
    "TR.F.CashDivPaidComStockBuybackNet(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0)",
    "TR.F.EPSBasicInclExordItemsComTot(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0)",
    "TR.F.EPSBasicExclExordItemsComTot(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0)",
    "TR.F.EPSDilInclExordItemsComTot(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0)",
    "TR.F.BookValuePerShr5YrCAGR(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FY0)",
    "TR.F.ReturnAvgComEqPct(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0)",
    "TR.F.LeveredFOCFPerShr(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,ReportingState=Orig, Period=FQ0)",
    "TR.EpsSmartEst(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, Period=FQ0)",
    "TR.NetProfitMean(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,Period=FQ0)",
    "TR.NetprofitSmartEst(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,Period=FQ0)",
    "TR.EBITDASmartEst(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,Period=FQ0)",
    "TR.RevenueMean(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,Period=FQ0)",
    "TR.F.LeveredFOCFPerShr(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ,Period=FQ0)",
    "TR.EPSSmartEstLastYrGrowth(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ)",
    "TR.EBITDASmartEstLastYrGrowth(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ)",
    "TR.RevenueSmartEstLastYrGrowth(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ)",
    "TR.EPSActSurprise(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, Period=FQ0)",
    "TR.RevenueActSurprise(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, Period=FQ0)",
    "TR.OperatingMarginPercent(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, Period=FQ0)",
    "TR.F.EBITDAMargPctTTM(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, Period=FQ0)",
    "TR.F.IncAfterTaxMargPct(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, Period=FQ0)",
    "TR.NetProfitMean(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, Period=FQ0)",
    "TR.F.NetDebtToEBITDATTM(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ)",
    "TR.F.NetDebttoTotEq(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, ReportingState=Orig, Period=FQ0)",
    "TR.F.EBITNetIntrCovRatio(SDate=2014-01-01, EDate=2024-05-30, Frq=FQ, ReportingState=Orig, Period=FQ0)",
    "TR.F.IntrExpnNetOfIntrInc(SDate=2014-01-01,EDate=2024-05-30,Period=FQ0,Frq=FQ,Methodology=StubLTM,ReportingState=Orig)",
    "TR.F.CurrRatio(SDate=2014-01-01,EDate=2024-05-30,Period=FQ0,Frq=FQ)",
    "TR.EpsSmartEst(SDate=2014-01-01,EDate=2024-05-30,Frq=FQ,Period=FQ0)",
    "TR.EPSActValue(SDate=2014-01-01,EDate=2024-05-30,Frq=FQ,Period=LTM)"
]
    


metrics_data_dicts = {}


for i in range(3):
    
    key_name = f"metrics_dict_{i}"
    logger.info(f"Iteration number {i}")
    
    metrics_data_dicts[key_name] = retrieve_and_process_fundamental_data(tickers=tickers, 
                                                         metric_codes=fields, 
                                                         ticker_partition_size=ticker_partition_size, 
                                                         metric_partition_size=metric_partition_size,
                                                         max_attempts=max_attempts)
pythonworkspace#technologypython apirefinitiv-data-librariesperformance
icon clock
10 |1500

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

Upvotes
Accepted
1.3k 3 2 4

Hi @alejandro.gonzalez ,

You're using the basic function rd.get_data() but you want to check the result with a granularity that this function doesn' t provide.

You should replace

metrics_data = rd.get_data(tickers, metric_codes)

with rd.content.fundamental_and_refrence API:

result = rd.content.fundamental_and_reference.Definition(universe=tickers, fields=metric_codes).get_data()
if result.errors:
    # errors is a not empty list of tuples (code, error_message)
    # => iterate on this list to detect error causes
    ...
else:
    # retrieve result as a DataFrame
    metrics_data = result.data.df


icon clock
10 |1500

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

Upvotes
82.3k 277 53 77

@alejandro.gonzalez

Thank you for reaching out to us.

As far as I know, I may relate to the server load. The server can cancel the request due to time out. You may try to reduce the number of fields or the date range (SDate and EDate) in reach request.

You can enable logging in the library to verify what the problem is by using the following code.

config = rd.get_config()
config.set_param("logs.transports.file.enabled", True)
config.set_param("logs.transports.file.name", "refinitiv-data-lib.log")
config.set_param("logs.level", "debug")
rd.open_session()

The refiniti-data-lib.log file will be created. You can check the log file for the issue.

icon clock
10 |1500

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

Write an Answer

Hint: Notify or tag a user in this post by typing @username.

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.