Create column based on datecolumn values in R - r

Question

I wrote a function that creates column based on a datetime column using parameters starting and ending dates, but I can't get it to work.
df is a data frame object.
create_gv <- function(df, s_ymd, e_ymd, char) {
df<-get(df)
for (i in (1:nrow(df))) {
ymd <- format(df[i,1],"%y%m%d")
if ((strptime(ymd,format = "%y%m%d") >= strptime(s_ymd,format = "%y%m%d") & strptime(ymd,format = "%y%m%d") <= strptime(e_ymd,format = "%y%m%d")) == TRUE) {
df$group_var[i]<-char
}
}
}
create_gv("example","171224","171224","D")
I get
> example
start_time group_var
1 2017-12-24 10:42:39 NA
2 2017-12-24 10:44:31 NA
3 2018-01-14 12:05:53 NA
4 2018-01-14 12:22:12 NA
Reproducible data frame named example here:
example <- structure(list(start_time = structure(c(1514112159, 1514112271, 1515931553, 1515932532), class = c("POSIXct", "POSIXt"), tzone = ""), group_var = c(NA, NA, NA, NA)), .Names = c("start_time", "group_var"), row.names = c(NA, -4L), class = "data.frame")
Desired output:
start_time group_var
1 2017-12-24 10:42:39 D
2 2017-12-24 10:44:31 D
3 2018-01-14 12:05:53 NA
4 2018-01-14 12:22:12 NA

Answer 1

From your description, my understanding is that you want to check if the date in a row is between the start and end date (which are scalars), and update the value of group_var accordingly.
The lubridate package provides a set of tools which allow to easily work with dates. In order to compare dates you don't need to format them. format only helps with the viewing of these dates. I have used the dplyr package which allows you to easily perform data transformations.
To solve the problem, we use the dplyr::mutate function which transforms a column by row, as a function of other columns. In this case, the date column in our dataset (start_time) is to compared with scalar start and end times in order to codify the group_var variable.
library(lubridate)
library(magrittr)
char <- "D"
# Randomly setting the start and end times for the purpose of the example. Any value can be passed to this.
s_ymd <- df$start_time[1] - 5000
e_ymd <- df$start_time[2] + 5000
df %>% dplyr::mutate(group_var = ifelse(start_time > s_ymd & start_time <
e_ymd,
char, NA)) -> df
df
To use a function directly, write:
create_gv <- function(start_time, s_ymd, e_ymd, char){
g_var <- ifelse(start_time > s_ymd & start_time < e_ymd,
char, NA)
return(g_var)
}
df %>% dplyr::mutate(group_var = create_gv(start_time, !!s_ymd, !!e_ymd,
!!char))
Here since s_ymd, e_ymd and char are scalars (i.e. not columns in the data frame), we need to unquote them. Note that the mutate function works on vectorized functions as desired.