I'd like to augment a zoo object with a variable which I could use to test for month changes. I'm sure there are more general ways to do this. Suggestions there would be great, but I'd like to understand why this simple approach fails. I'd feel better if I understood what I'm missing here ;-)
e.g. for a zoo object
library(zoo)
tz <- structure(c(7L, 7L, 1L, 6L, 0L, 9L, 0L, 1L, 6L, 0L, 3L, 3L, 5L,
0L, 8L, 2L, 0L, 3L, 2L, 5L, 2L, 3L, 4L, 7L, 8L, 9L, 0L, 1L, 4L,
5L, 6L, 7L, 8L, 2L, 3L, 4L, 5L, 8L, 9L, 0L), .Dim = c(20L, 2L
), .Dimnames = list(NULL, c("x", "y")), index = structure(c(13880,
13881, 13913, 13916, 13946, 13947, 13948, 13980, 13983, 13984,
13985, 14016, 14048, 14082, 14083, 14115, 14147, 14180, 14212,
14243), class = "Date"), class = "zoo")
Add a year/month variable using as.yearmon() seems easy enough. If I were in a data frame this would yield a fine character variable, but in zoo tragedy ensues if you forget to wrap in as.numeric()
tz$yrmo <- as.numeric(as.yearmon(index(tstz)))
> head(tz)
x y yrmo
2008-01-02 7 2 2008.000
2008-01-03 7 3 2008.000
2008-02-04 1 4 2008.083
2008-02-07 6 7 2008.083
2008-03-08 0 8 2008.167
2008-03-09 9 9 2008.167
This looks great and I can compare data elements successfully
(tz$x[6] != tz$y[6])
2008-03-09
FALSE
but why do I get this result when I compare the year/month variable?
> (tz$yrmo[2] != tz$yrmo[1])
Data:
logical(0)
Index:
character(0)
and why does testing the yearmon or data items with identical() fail in this way? (both should be true)
> identical(tz$yrmo[2] , tz$yrmo[1])
[1] FALSE
> identical(tz$x[2] , tz$x[1])
[1] FALSE
Am I just playing with fire in using yearmon() which creates an index class in zoo? Should I switch to something like Dirk Eddelbuettel's 'turning a date into a monthnumber'? Number of months between two dates