For a deeper look into our Eikon Data API, look into:
Overview | Quickstart | Documentation | Downloads | Tutorials | Articles
Time of daily rollover for historical daily fx data?
I would like to know
A) at what time historical timeseries for fx data roll over when I run a python request. What is the default timezone and time of the historic data rollover for fx data?
B) can I request in python a specific timezone or time when requesting historical daily fx data? For example I want the daily close to be EST 5pm or UTC midnight for historical fx time series requests. Bloomberg offers that with parameter overrides. How can I do this with the python api via Eikon?
@HFTVOL Hi Matt ,
To summarise our conversation
Rules for all currencies are the same.
In OPEN you will see first tick after 5 PM New York time (21:00 GMT Summer time)/22:00 GMT (Winter Time) except of Monday’s session, where we capture first tick placed after 07:00 Wellington time (19:00 GMT Sunday at Summer/18:00 GMT at Winter). Just as example with AUD= , tick at 20:59:59 GMT (16:59:59 NY time) on 24th May is marked as Close price, while next tick 21:00:01 (17:00:01 NY time) is marked as 25th May Open. If there will be any ticks placed prior to 07:00 Wellington time on Monday – they will be not considered as open price and overwritten by first tick after 07:00. 
For close we capture last available tick prior to 5 PM New York time on Monday-Friday
Also , as alternative we do have regional snaps for historical data – “TR.” Fields I can suggest following one:
TR.ASIACLOSEBIDPRICE – snaps last tick available prior to 08:00 GMT on Monday - Friday
TR.EUROPECLOSEBIDPRICE – snaps last tick available prior to 17:00 London Time on Monday- Friday
TR.AMERICACLOSEBIDPRICE – snaps last tick available prior to 17:00 New York time (matching with close)
Just one reminder – if you collect regional snaps for illiquid currencies (let’s say ARS=) and there was no update during region time - field will be populated with previous value.
One more important detail – if you try to retrieve timestamp for above fields – it return 00:00 – this may be a bit misleading, but this behaviour is observed because we expect to update TR fields only once a day. We may expect TR fields to be updated with current day value after market close (5 PM New York time).
Also, there are RICs allowing to collect timeseries of hourly snaps, their construction is EUR07H= where 07 is the hour of snap (in GMT time zone whole year). It will update once a day and snap value from main RIC (EUR=) at the mentioned time.
Thanks
Marek
A) Different asset classes and markets have what is called a Closing Run which clears down the session data and archives, and readies it for the next session. These Closing Runs will vary by geography, exchange/TC venue, asset class. For more details please raise a content query where they can furnish you details of closing run for particular RICs you are interested in. Im not an FX specialist so im not so sure about the details of that asset class.
B) All real-time data on our platform is timestamped in GMT - once data is on our platform we sometimes put a gmt offset where appropriate for historical timeseries or for example news. With python you can use the excellent pytz package. Full examples are contained there and it also contains a full list of timezones etc.
I hope this can help.
No, this unfortunately does not help at all. Why are you talking about Realtime data? I asked about historical time series data.
And why post an answer when you say yourself you are not familiar with the issue? With fx I mean fx symbols such as EUR=.
I think my question is simple: when pulling historical time series for fx symbols, in what time zone and at what time are the daily rollovers? At what time does each day's close occur and at what time does the open of each bar occur? I am specifically asking about fx data.
@HFTVOL Historical timeseries data is just an archive of our streaming data. I was trying to explain to you the process as you were asking about when does the history update. I don't know the exact times so I have opened a content query on your behalf and they will be contacting you shortly with the details you requested. Your ticket number is 09933169. Let me know how it goes.
@jason.ramchandani, I could not get any useful answer from customer service.
I am surprised why my question is seemingly difficult to address. All cash fx historical data should be identically stored. For daily data at what exact time and of which timezone do the daily closes pertain to?
@HFTVOL Really sorry about the poor response from customer service - I have sent you an email with an update on what I have discovered. I will await confirmation from the FX content team before posting the answer here.
Hello! Regarding the question about time series, is there a way to obtain the tick by tick in Python? I have the formula but is just for Daily.
df, err = ek.get_data(
instruments = ['MXN='],
fields = ['TR.ASKPRICE(SDate=2011-01-03,EDate=2021-05-28,Frq=D)']
)
display(df)