[R-SIG-Finance] IBrokers

[R P Herrold]

On Mon, 1 Dec 2008, Arno gaboury wrote:

> My only problem is to be able to download historical data 
> (250 days) with a customized closing time (e.g. 5:00 PM 
> GMT), not the regular closing price given by exchanges. 
> Untill now, after weeks & weeks of browsing the web, I 
> couldn't find any suitable solution. I thought IBrokers 
> package could be one.

Jeffrey A. Ryan (author of IBrokers) did not supply a worked 
example, and I think I have properly stated it later in this 
email.

Ad hoc history retrieval intervals are very doable with IB's 
history farms (I do not comment yet on the R connection tool 
plugin) -- 1700 GMT is either 1200 or 1300 US Eastern time 
[varies on 'daylight savings time' offset], and so one would 
pull a query on that time mark

A sample script to do such retrievals was published some 
months ago at:
     http://www.trading-shim.org/pipermail/ts-general/2008-March/000168.html

Obviously one would alter the query time parameter to just:
 	"12:00:00"   or "13:00:00"
in the 'time' array; and update 'syms' to taste;  the Ruby 
script handles 'walking' through successive array element 
conbinations



In like package, the IBrokers package has
 	reqHistoricalData
and I would expect such a parallel solution to look like:

(taken from the documentation at: 
http://cran.r-project.org/web/packages/IBrokers/IBrokers.pdf)

tws <- twsConnect()
contract <- twsContract("AAPL","STK","SMART","ISLAND",
                         "","0.0","USD","","","",NULL,NULL,"0")
reqHistoricalData(tws,
                   contract,
                   '20081201 12:00:00',
                   barSize = "1 min",
                   duration = "1 M",
                   useRTH = "1",
                   whatToShow = "TRADES",
                   time.format = "1",
                   verbose = TRUE,
                   tickerId = "1",
                   eventHistoricalData,
                   file)

# endDateTime argument must be of the form ’CCYYMMDD HH:MM:SS TZ’.

and we would simply step through successive "CCYYMMDD" days 
for the data;  IB does not keep the finest granularity history 
data around for a full year, but one minute data is usually 
available.

-- Russ herrold