7

I'm trying to convert a record into a date and time format using the strptime function. However, I'm not sure why I'm getting the error:

number of items to replace is not a multiple of replacement length.

I tried to check the length of the record using the length function but both have the same length. 

data <- DT
head(data[6])
#                column
# 1 2014-12-22 23:53:48
# 2 2014-12-22 23:20:34
# 3 2014-12-22 23:20:30
# 4 2014-12-22 23:20:16
# 5 2014-12-22 23:20:07
# 6 2014-12-22 23:05:49

data[,6] <- as.character(data[,6])

temp_file <- matrix(0,nrow=nrow(data))

temp_file[1] <- strptime(data[1, 6],"%F %T")
# Warning message:
# In temp_file[1] <- strptime(data[1, 6], "%F %T") :
#   number of items to replace is not a multiple of replacement length

length(temp_file[1])
# [1] 1

length(data[1,6])
# [1] 1

length(strptime(data[1, 6], "%F %T") )
# [1] 1

Any help is greatly appreciated.

Thanks!

Artem
  • 3,304
  • 3
  • 18
  • 41
dixi
  • 680
  • 1
  • 13
  • 27
  • 6
    That is actually a good question. This isn't an error, rather a warning, but the result you get is wrong, so you could consider it as an error. The reason this is happening is because of the definition of a `matrix` in R, which can only get atomic vectors. When you are trying to pass `strptime` to the matrix, it's class is `"POSIXlt" "POSIXt"`, thus it is unclasses it and thus returns a list of its attributes (which length is larger than 1), i.e., `unclass(strptime(data[1,1],"%F %T"))`. The first value is `48` seconds. This is exactly what you have in `temp_file[1]` now. – David Arenburg Jan 07 '15 at 13:17
  • 2
    Thus, either use `data.frame` instead of a `matrix` or try to convert your vector to atomic, for example: `temp_file[1] <- as.character(strptime(data[1,1],"%F %T"))` – David Arenburg Jan 07 '15 at 13:20

1 Answers1

0

You can transform character vector into datetime format using ymd_hms function of lubridate package:

library(lubridate)

# data frame simulation
structure(list(X1 = c(1, 1, 1, 1, 1, 1), X1.1 = c(1, 1, 1, 1, 1, 1), 
    X1.2 = c(1, 1, 1, 1, 1, 1), X1.3 = c(1, 1, 1, 1, 1, 1), 
    X1.4 = c(1, 1, 1, 1, 1, 1), date_time_char = c("2014-12-22 23:53:48", 
    "2014-12-22 23:20:34", "2014-12-22 23:20:30", "2014-12-22 23:20:16", 
    "2014-12-22 23:20:07", "2014-12-22 23:05:49")), class = "data.frame", row.names = c(NA, -6L))

# transform from character to datetime
data$date_time <- ymd_hms(data[, 6])
data[, 7]

Output:

[1] "2014-12-22 23:53:48 UTC" "2014-12-22 23:20:34 UTC" "2014-12-22 23:20:30 UTC" "2014-12-22 23:20:16 UTC"
[5] "2014-12-22 23:20:07 UTC" "2014-12-22 23:05:49 UTC"

N.B. Really nice comment by David Arenburg:

That is actually a good question. This isn't an error, rather a warning, but the result you get is wrong, so you could consider it as an error. The reason this is happening is because of the definition of a matrix in R, which can only get atomic vectors. When you are trying to pass strptime to the matrix, it's class is "POSIXlt" "POSIXt", thus it is unclasses it and thus returns a list of its attributes (which length is larger than 1), i.e., unclass(strptime(data[1,1],"%F %T")). The first value is 48 seconds. This is exactly what you have in temp_file[1] now.

Artem
  • 3,304
  • 3
  • 18
  • 41