I have a dataframe for which I've calculated and added a difftime
column:
name amount 1st_date 2nd_date days_out
JEAN 318.5 1971-02-16 1972-11-27 650 days
GREGORY 1518.5 <NA> <NA> NA days
JOHN 318.5 <NA> <NA> NA days
EDWARD 318.5 <NA> <NA> NA days
WALTER 518.5 1971-07-06 1975-03-14 1347 days
BARRY 1518.5 1971-11-09 1972-02-09 92 days
LARRY 518.5 1971-09-08 1972-02-09 154 days
HARRY 318.5 1971-09-16 1972-02-09 146 days
GARRY 1018.5 1971-10-26 1972-02-09 106 days
I want to break it out and take subtotals where days_out is 0-60, 61-90, 91-120, 121-180.
For some reason I can't even reliably write bracket notation. I would expect
members[members$days_out<=120, ] to show just Barry and Garry, but I get a whole lot of lines like:
NA.1095 <NA> NA <NA> <NA> NA days
NA.1096 <NA> NA <NA> <NA> NA days
NA.1097 <NA> NA <NA> <NA> NA days
Those don't exist in the original data. There's no one without a name. What am I doing wrong here?
This is standard behavior for <
and other relational operators: when asked to evaluate whether NA
is less than (or greater than, or equal to, or ...) some other number, they return NA
, rather than TRUE
or FALSE
.
Here's an example that should make clear what is going on and point to a simple fix.
x <- c(1, 2, NA, 4, 5)
x[x < 3]
# [1] 1 2 NA
x[x < 3 & !is.na(x)]
# [1] 1 2
To see why all of those rows indexed by NA
's have row.names like NA.1095
, NA.1096
, and so on, try this:
data.frame(a=1:2, b=1:2)[rep(NA, 5),]
# a b
# NA NA NA
# NA.1 NA NA
# NA.2 NA NA
# NA.3 NA NA
# NA.4 NA NA
If you are working at the console the subset
function does not have that annoying 'feature' which is actually due to the behavior of [
more than to the relational operators.
subset(members, days_out <= 120)
If you are programming, then you can use which
or Josh's conjunction with & is.na(.)
that which
does behind "the scenes":
members[ which(members$days_out <= 120), ]
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.