CODEWITHSUNDEEP
rxtsposixctposixlt
N_R
N_R

Reputation: 31

R - xts object encountering character string is not in a standard unambiguous format

I was trying to replicate the momentum strategy by rbresearch with S&P 500 companies. However it turned out that getSymbols was not able to grab price data for all tickers and I only got around 300 stocks by using: 

getSymbols(symbols, src='yahoo', index.class=c("POSIXt","POSIXct"), from='2000-01-01', to = '2015-12-31')

I found this thread discussing about this issue and it might be coming from the source "chart.yahoo.com". Therefore I adopted the method suggested by jlhoward and it seemed to work perfectly without warnings.

UPDATE: After checking quantmod package, I found it is also grabbing data from ichart.yahoo.com now.

However when it comes to calculating the monthly close, an error read like, 

Error in try.xts(x) : 
  Error in as.POSIXlt.character(x, tz, ...) :   character string is not in a standard unambiguous format
Called from: try.xts(x)

Here are my questions and what I don't understand:

1) Is there a possible way to direct getSymbols to "ichart.yahoo.com" which gives more reliable quotes?

2) I have converted all symbols to xts objects so why the error was called from try.xts?

3) I presume the problem is with the dates, which are not in ISO 8601 format. But I haven't found a way to convert the dates to POSIXct since the dates are only round to days.

4) Any comments to the codes are much appreciated.

I attached with the codes and the sp500components.csv could be downloaded from here.

Thanks for your time and kind help! All the best. 

library(quantstrat)
library(FinancialInstrument)
library(TTR)

symbols <- read.table("sp500components.csv", header = FALSE, sep = ",")$V1
symbols <- as.character(symbols) 

currency("USD")
stock(symbols, currency="USD",multiplier=1)

MonthlyAd <- function(x){
  # Converts daily data to monthly and returns only the monthly close 
  # Note: only used with Yahoo Finance data so far
  # Thanks to Joshua Ulrich for the Monthly Ad function
  # 
  # args:
  #   x = daily price data from Yahoo Finance
  #
  # Returns:
  #   xts object with the monthly adjusted close prices

  sym <- sub("\\..*$", "", names(x)[1])
  Ad(to.monthly(x, indexAt = 'lastof', drop.time = TRUE, name = sym))
}

symEnv <- new.env()

f <- function(x) {
  uri    <- "http://ichart.yahoo.com/table.csv"
  symbol <- paste("s",x,sep="=")
  from   <- "a=1&b=1&c=2000"
  to     <- "d=31&e=12&f=2015"
  period <- "g=d"
  ignore <- "ignore=.csv"
  query  <- paste(symbol,from,to,period,ignore,sep="&")
  url    <- paste(uri,query,sep="?")
  try(assign(x,read.csv(url),envir=symEnv))
}

lapply(symbols,f) 

ts <- function(x) {
    x["Date"] <- as.Date.character(x[["Date"]], "%Y-%m-%d")
    x <- xts(x[,-1], order.by = x$Date)
    colnames(x) <- gsub("Adj", "Adjusted", colnames(x))
    assign(symbol, x)
}

eapply(symEnv, ts) 

symbols.close <- do.call(merge, eapply(symEnv, MonthlyAd))

Upvotes: 1

Views: 1027

Answers (1)

Joshua Ulrich
Joshua Ulrich

Reputation: 176648

Since you did not provide a reproducible example of the cause of the error, I have to make some guesses. So I assume you're calling MonthlyAd on the objects in symEnv.

The object created by your f function does not have the characteristics that to.monthly expects. Namely, the data.frame it creates does not have row names that represent the datetimes for each observation.

There's no problem if you call MonthlyAd on the objects created by eapply(symEnv, ts).

symList <- eapply(symEnv, ts)
symListMonthly <- lapply(symList, MonthlyAd)

There's also no problem if you just use getSymbols an the env argument. For example:

symEnv <- new.env()
getSymbols("SPY;IWM", env = symEnv)
symListMonthly <- eapply(symEnv, MonthlyAd)

Upvotes: 1

Related Questions