问题
I've been going through Guy's quantstrat lecture (link below) and after repeatedly attempting to re-execute the code, I'm getting a few initial errors that are preventing most of the subsequent code in the lecture from functioning.
Here is the code (copied from the lecture with very minor re-arrangements):
rm(list=ls(all=TRUE)) #added this to delete memory
library(quantstrat)
library(blotter) #added this hoping it would rectify the errors
library(FinancialInstrument) #added this hoping it would rectify the errors
# initialize portfolio, accounts and orders
qs.strategy <- "qsFaber"
initPortf(qs.strategy, 'SPY', initDate = '1997-12-31')
initAcct(qs.strategy, portfolios = qs.strategy, initDate = '1997-12-31', initEq= 1e6)
Here are the errors I am getting:
1)
> initPortf(qs.strategy, 'SPY', initDate = '1997-12-31')
Error in exists(paste("portfolio", name, sep = "."), envir = .blotter, :
object '.blotter' not found
2)
> initAcct(qs.strategy, portfolios = qs.strategy, initDate = '1997-12-31', initEq= 1e6)
Error in exists(paste("account", name, sep = "."), envir = .blotter, inherits = TRUE) :
object '.blotter' not found
I had to directly download blotter as I am using Windows 64 bit, but despite copying the code from the lecture, I am unsure why I am getting those errors. My search efforts have indicated that a portion of blotter evolved into the FinancialInstrument package, but even after clearing memory and loading FinancialInstruments I am still getting the same error.
Any help would be highly appreciated.
LINK to lecture: http://www.r-programming.org/files/quantstrat-I.pdf
回答1:
The sheets by Guy Yollin are excellent learning material, but unfortunately they are somewhat outdated (2011). Many changes have been made to blotter, quantstrat and other packages over the last 2 years, and much of the code in Guy's sheets will no longer run as such.
As far as the quantstrat package is concerned, you may want to take a look at the sheets from the R/Finance 2013 conference in Chicago; you can get a copy at http://www.rinfinance.com/agenda/2013/workshop/Humme+Peterson.pdf.
UPDATE: Guy Yollin has updated his slides to the latest quantstrat as of August 2013, they are available here http://www.r-programming.org/papers
回答2:
The blotter and quantstrat packages store things in the .GlobalEnv (which is one reason they're not on CRAN.) When you run rm(list=ls(all=TRUE)), you are removing things, that those packages expect to be able to find in your workspace. In order for everything to work, you have to put a couple environments back in your globalenv(). After running these two lines of code, I think your code will work.
.blotter <- new.env()
.strategy <- new.env()
In the past, FinancialInstrument used to create a .instrument environment in the .GlobalEnv (and later expect it to exist). A couple years ago, I changed it so that .instrument is now stored in the FinancialInstrument namespace. Since that change came after Guy's slides, the code is not compatible. Slides 14-15 should be changed to
currency("USD")
getInstrument("USD")
stock("SPY", "USD")
getInstrument("SPY")
Or to more closely follow his original code,
get("USD", envir=FinancialInstrument:::.instrument)
get("SPY", envir=FinancialInstrument:::.instrument)
By storing package level objects in the package's namespace, the user is free to remove everything from the globalenv() without breaking any of the package's code.
来源：https://stackoverflow.com/questions/17120496/guy-yollins-quantstrat-i-lecture-issue