[R-SIG-Finance] how to grow XTS series in R dynamically ? And Quickly!

[Vladimir Morozov]

did some googling on my question
general wisdom on the internet seems to be - it's SLOW to grow R lists in a
loop. Must Pre-allocate First.

Question: is there a feature in XTS that allows to pre-allocate space for
growing XTS series?
thanks again!

On Fri, Sep 6, 2019 at 11:51 AM Vladimir Morozov <vmorozov2006 using gmail.com>
wrote:

> Dear xts experts
>
> I use R and XTS to store dynamically growing price time-series for
> currency pair rates (e.g. time series of EUR/USD, EUR/JPY, etc growing with
> each new incoming market price - say, one per second).
>
> I want to speed up my R function that generates and rbinds a few new rows
> to multiple Global XTS objects. This function us called often (sometimes
> once per second) and takes a lot of time.
>
> Specifically, I have a function add.newRate.EUR(POSIXct Time, float Rate)
> This function performs some time-consuming manipulations, then modifies
> (global objects) XTS series1 through XTS seriesN. Specifically, I rbind
> some new rows to each of series1 through seriesN in global environment.
> (specific example - a new Rate for EUR/USD comes into the function. This
> function needs to rbind new point to EUR.USD xts object, but it also needs
> to calculate new value for cross-currency pairs, like EUR/JPY, EUR/GPB,
> EUR/CHF, EUR/CAD, EUR/AUD, EUR/NZD, and rbind new point to those XTS series
> as well)
>
> it all happens quite slowly in R. How do I accelerate it?
>
> To avoid copying XTS series between the function and the global
> environment, the function attempts to modify the seriesX in the global
> environment, and not return XTS on function return.
> The structure of my project is such that my other tasks in R need to
> access these XTS series1 through seriesN in Global Environment.
> Still, this whole set-up is rather slow.
>
> Do you think using Rcpp to move cross-currency rate calculations to C++
> could help?
> E.g. I want to write add.newRate.EUR.RCPP(int Time, float Rate) in C++. I
> think it will perform internal manipulations of time T and value X a little
> faster.
>
> Second idea for speed up:  doing rbind in C++ using package RcppXts.
> I plan to use the following function in RcppXts package:
>     function("xtsRbind",
>              &xtsRbind,
>              List::create(Named("x"), Named("y"), Named("dup")),
>              "Combine two xts objects row-wise");
> Then I want to use Rcpp function assign( name, x ) to assign the created
> object back to GlobalEnvironment.
>
> Do you think the above way is a good way to grow dynamical XTS series inside
> Rcpp?
> Do you think using Rcpp and RcppXts will provide significant acceleration
> compared to pure R ?
>
> Thank you
> Vlad
>
>
>

	[[alternative HTML version deleted]]