Merging price time series in xts when securities show no data - r

Question

I am querying a security DB. Price series are in xts and for some there might be no data (for the chosen window). Actual time series can be simulated as follows:
require(xts)
## Simulated time series
price=function(){
x=floor(runif(1,1,4))
xts(round(rnorm(x,5),3), Sys.Date()+1:x)
}
## Sample tickers
(tick1=setNames(price(), "tick1"))
# tick1
# 2014-04-20 5.829
# 2014-04-21 6.061
# 2014-04-22 5.813
(tick2=setNames(price(), "tick2"))
# tick2
# 2014-04-20 6.458
# 2014-04-21 5.373
(tick3=xts(data.frame(tick3=numeric()), as.Date(numeric()))) # Security showing no data
# tick3
## ...
## tickn
No need to mention that I don't know in advance which security will show no data.
If I merge the prices in a single xts object, merge.xts completely removes from the output the empty security(ies):
(port=merge(tick1, tick2, tick3))
# tick1 tick2
# 2014-04-20 5.829 6.458
# 2014-04-21 6.061 5.373
# 2014-04-22 5.813 NA
Instead I would like to keep trace of them, therefore printing an output similar to:
(cbind(port, tick3=NA))
# tick1 tick2 tick3
# 2014-04-20 5.829 6.458 NA
# 2014-04-21 6.061 5.373 NA
# 2014-04-22 5.813 NA NA
One possible solution is:
port=list(tick1, tick2, tick3) # ... tickn
port.m=lapply(port, function(sec){
if(nrow(sec)==0) sec= xts(matrix(NA, dimnames=dimnames(tick3)), Sys.Date())
sec
})
(port.m=do.call('merge', port.m))
# tick1 tick2 tick3
# 2014-04-19 NA NA NA
# 2014-04-20 5.829 6.458 NA
# 2014-04-21 6.061 5.373 NA
# 2014-04-22 5.813 NA NA
if(all(is.na(port.m[Sys.Date()])))
(port.m=port.m[time(port.m)!=Sys.Date()])
# tick1 tick2 tick3
# 2014-04-20 5.829 6.458 NA
# 2014-04-21 6.061 5.373 NA
# 2014-04-22 5.813 NA NA
Is it possible to find a smarter solution?

Answer 1

You are making two mistakes here:
First: You need to use a vector of non-zero length. See this:
length(integer())
length(NA)
Second: For merge to work, the xts object indices have to match some where.
e.g. something like this will work:
require(xts)
x=xts(1:4, Sys.Date()+1:4)
v=xts(NA, Sys.Date()+1)
(m=merge.xts(x,v))
Here the starting index matches, and remaining indices are filled up.
If you want to be very particular, you could probably try something like this:
v=xts(rep(NA,4), Sys.Date()+1:4)
Hope this helps!!