I use the LSEG Data Library for Python (lseg.data) against a Workspace desktop session, calling ld.get_data. A typical call:
ld.get_data(universe=<20 equity RICs>, fields=["TR.CLOSEPRICE.date", "TR.CLOSEPRICE", "TR.Volume"], parameters={"SDate": 0, "EDate": -1200, "FRQ": "d"})
That is up to ~24,000 rows. In a test of 55 such calls, 29 came back partial with no error: HTTP 200, no exception, no warning. One returned 1,201 rows — 19 of its 20 RICs were empty. Re-running the same call usually returns everything.
- Is there a documented per-request limit (rows, data points, cells or response size) for ld.get_data on a desktop session? What is the number, and is it different for ld.get_history or the historical-pricing endpoints?
- When a request hits that limit or a throttle, is the response truncated silently? Is there any flag, status, warning, header or error object that marks a response as partial, and how do I read it from Python?
- What are the request limits for a Workspace desktop session (per second, per minute, per day, concurrent requests)?
- To pull ~1,100 equities x 1,200 daily closes and volumes once, then 1-5 new days per instrument every day, which call, chunk size and pause between calls do you recommend?
Could you cite the documentation page for each answer, and say plainly when something is not documented?
