0

I'm trying to read a set of csv datasets into R as zoo time series. I've managed to convert the data into data frames with the index columns as dates and the rest as numeric vectors. I now need to convert these data frame into zoo. Here's the code I'm using:

for(x in 1:length(files)){
    for(n in 1:length(files[[x]]){
        files[[x]][[n]]<-unique(files[[x]][[n]]) #removes duplicate rows
        }
    zfiles[[x]]<-lapply(files[[x]],function(n) read.zoo(n))
    }

Upon running it I get this error: Error in read.zoo(n) : index has bad entry at data row 4

I've tried randomly converting individual datasets within files and i do not get any error, so I'm assuming there are certain problem sets which cause the issue. Is there any way to remove the bad entries before converting to zoo?

Thanks

user3294195
  • 1,748
  • 1
  • 19
  • 36

2 Answers2

1

Turns out the bad entry rows were rows with NA. Fixed it using complete.cases, as per this link: Remove rows with NAs (missing values) in data.frame

Community
  • 1
  • 1
user3294195
  • 1,748
  • 1
  • 19
  • 36
0

make sure your dates column is in the format of a POSIXct, then use 

zoo(<vector of numeric>, order.by=<vector of POSIXct>)
RockScience
  • 17,932
  • 26
  • 89
  • 125
  • Thanks, I've tried this, but got the following error: ` Error in .subset2(x, i, exact = exact) : recursive indexing failed at level 2` – user3294195 Apr 09 '14 at 06:53
  • Converted the dates to POSIXct and used that to order the zoo series. – user3294195 Apr 09 '14 at 06:56
  • Also, each data frame contains multiple numeric vector columns. – user3294195 Apr 09 '14 at 06:59
  • 1
    @user3294195 as long as you don't provide a reproducible example, nobody will be able to help you. http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example – RockScience Apr 09 '14 at 07:55