I am trying to pull historical economic indicator with ek.get_data. While I was able to obtain the time series with ek.get_timeseries, the date returned was in the period which the indicator was responsible for (ie If I'm looking at the CPI released on Oct 12 2023, the data reported was responsible for 2023-09-30. I saw a related problem here and would like to have your insights if there's anyway to get_timeseries to return the (original release date) that I'm looking for.
df=ek.get_timeseries(['USCPNY=ECI'],fields=['TIMESTAMP', 'TRADE_DATE','VALUE'],calendar='native',start_date='2007-03-01',interval='monthly')#,fields=['DSPLY_NMLL','RELEVANCE','ACT_REL_ST','ECI_ACT_DT','ACT_VAL_NS','UNIT_PREFX','FCAST_PRD','RPT_UNITS','ECON_ACT','ECON_PRIOR','RTR_POLL','FCAST_SEST'])
display(df)
Figure1. Output from notebook
Figure2. Desired output should have date: October 12 2023.
Another approach I attempted: use ek.get_data, which I found the CF_DATE to correspond to the date of release, but in this case ek.get_data returns a single point rather than a dataframe/timeseries.
df, err = ek.get_data(
instruments = ['USCPNY=ECI'],
fields = ['CF_TIME','CF_DATE',
'TR.IndicatorLastObservationDate',
'TR.IndicatorStartDate',
'TR.IndicatorType',
'TR.IndicatorSeasonalAdj',
'TR.IndicatorBaseYear',
'TR.IndicatorPeriodicity',
'ACT_STA_DT',
'ACT_END_DT',
'ACT_VAL_NS',
'ACT_END_DT',
'ACT_VAL_NS',
'NDOR_1',
'TRADE_DATE',
'REF_PRD',
'ECON_ACT',
'GN_TXT16_4'
],parameters={"SDate":"2007-01-01",'EDate':"2023-10-14