get_history bulk request

valstar

Can somebody help me make this request more efficient, so that it will consume less requests?

Goal: i need daily timeseries for several instruments (options) for some fields.

In the past, way before LSEG, when it was still called get_timeseries instead of get_history, i was able to retrieve timeseries for multiple instruments at once, and it would only count towards ONE request. This was great. Now however, it counts towards one request PER instrument. I obviously don't have enough request limit to do that.

import lseg.data as rd

with rd.open_session() as lseg_session:

df = rd.get_history(

universe=['BL219300C6','BL219400C6'], #.. many more instruments.. ~10.000 option contracts

fields=['SETTLE', 'IMP_VOLT', 'OPINT_1', 'DELTA', 'GAMMA', 'THETA', 'VEGA', 'RHO'],

interval='1d',

start=SCRAPE_DATE.date() - timedelta(days=3) , #3 days ago

end=SCRAPE_DATE.date() - timedelta(days=1) # yesterday

)

i've been playing with get_data as well, but couldnt get a date.

Jirapongse

@valstar

Thank you for reaching out to us.

When requesting interday historical data for real-time fields that do not have the TR. prefix, the library uses the Historical Pricing API to retrieve the data. This API supports only one item per request.

You may find the TR.xxx fields that can provide the required data.

You can use the Data Item Browser tool to search for fields and parameters. For example:

df = ld.get_data(
    universe = ['BL219300C6','BL219400C6'],
    fields = [
        
        'TR.SETTLEMENTPRICE.Date','TR.SETTLEMENTPRICE',
        'TR.IMPLIEDVOLATILITY',
        'TR.OPENINTEREST',
        'TR.Delta',
        'TR.Gamma',
        'TR.Vega',
        'TR.RHO'],
    parameters = {'SDate': '2025-11-02', 'EDate': '2025-11-04', 'FRQ': 'D'})
df

However, you should not add all 10.000 option contracts into the universe parameter at once. Please check the limitations of the get_data method from the Usage and Limits Guideline.

Moreover, it is better to use the rebranded LSEG Data Library for Python instead. The examples are on GitHub.