hi,

I want to calculate the correlation between an index and an etf, but there is a mistace in my code and I don't know how remedy the defect.

this is my code:
library("quantmod")
library("fPortfolio")
library("RBloomberg")
conn <- blpConnect()

start.date = as.POSIXct("2012-10-01")
end.date = as.POSIXct(Sys.Date())
periodicity <- "DAILY"
price <- NULL


etf <- c("GLD US Equity")


etf_x <- etf_bla

ticker <- c("INDU Index")

get_index_value <- function(index, date) {
for(r in 1:nrow(index)) {
if(index[r,1] == date)
{
return(as.numeric(index[r,2]))
}
}
return(NA)
}

calculate_missing_days <- function(revenues) {

print("calculating missing days ...")

# remove all NA's
revenues_no_na <- revenues[which(!is.na(revenues))]

n <- length(revenues_no_na)

v <- var(revenues_no_na)
m <- mean(revenues_no_na)


sumvar <- 0
for(i in 1:length(revenues_no_na))
{
sumvar <- sumvar + (revenues_no_na[i] - m) ^ 2
}

x = sqrt((v * n - sumvar)) + m

print(sprintf("replacing NAs with x=%f ...", x))

revenues[which(is.na(revenues))] = x


return(revenues)
}


calc_correlation_to_index <- function(etf, ticker) {

load_index_data <- function(etf) {
for(i in 1:length(etf)) {
download <- bdh(conn, etf[i], "PX_LAST", start.date, end.date, option_names = "periodicitySelection", option_values = periodicity)
a <- matrix(data=download$PX_LAST, ncol=1)
rownames(a) <- unique(download$date)
colnames(a) <- etf[i]
price <- merge.xts(price,as.xts(a))
}
}

etf_data <- load_index_data # r cannot find "etf_data"
combined_data <- matrix(, ncol=2, nrow=0)

#print(etf_data)
#print(index_data)

#first walk through all etf data and find the fitting index data to it
for(i in 1:nrow(etf_data)) {

date <- time(etf_data[i])
date_string <- format(date, format="%Y-%m-%d")

row <- matrix(, ncol=2)
#row[1,1] = date_string
row[1,1] = as.numeric(etf_data[i,6])
row[1,2] = get_index_value(index_data, date_string)

combined_data <- rbind(combined_data, row)
}
#print(combined_data)



#print(cor(combined_data[,1], combined_data[,2]))

# calculate revenues now
combined_revenues <- matrix(,ncol=2, nrow=nrow(combined_data)-1)

# calcuclate revenues for etf data
combined_revenues[,1] <- (as.numeric(combined_data[2:nrow(combined_data),1]) / as.numeric(combined_data[1:nrow(combined_data)-1,1])) - 1
# calculate revenues for index data
combined_revenues[,2] <- (as.numeric(combined_data[2:nrow(combined_data),2]) / as.numeric(combined_data[1:nrow(combined_data)-1,2])) - 1

#print(combined_revenues)

sumnas = length(which(is.na(combined_revenues[,2])))

combined_revenues[,2] = calculate_missing_days(combined_revenues[,2])
#print(combined_revenues)

print(sprintf("calculating correlation of %s ...", etf))
correlation <- cor(combined_revenues[,1], combined_revenues[,2])

ret <- matrix(,ncol=3, nrow=1)
ret[1,1] <- etf
ret[1,2] <- correlation
ret[1,3] <- sumnas

return(ret)
}

print_etf_highscore <- function(indexfile, etf_list) {
index_data <- load_index_data(indexfile)
etf_highscore <- matrix(, ncol=3, nrow=0)
for(etf in etf_list) {
print(sprintf("############### %s ###############", etf))
row <- calc_correlation_to_index(etf, ticker)
etf_highscore <- rbind(etf_highscore, row)
}

print(sprintf("############### HIGHSCORE ###############"))
print(etf_highscore[rev(order(etf_highscore[,2])),])
}

print_etf_highscore(ticker, etf)

I think the main problem is, that R can't find "etf_data". And as a result R tells me, that "1:nrow(etf_data)" has the length zero. Does sombody how to fix this?

I would be deeply greatfull.

brigit