How to split column containing date and time when null values are present in R?-CodePudding

I have a column that has data as such

df <- data.frame(request_time = c("2020-12-31 13:05:00", NULL, "2020-11-14 02:04:01")

I want to split the request_time column to extract just the the date. Hoping to have a new column called request_date.

I was trying to do the following:

df$request_date <- as.Date(df$request_time)

But this returns an error saying the "character string is not in a standard unambioguous format" Im assuming dates due to the NULLS present. How can I get past this?

CodePudding user response：

We could use as_date function together with ymd from lubridate package:

library(dplyr)
library(lubridate)
df %>% 
  mutate(request_time = ymd(as_date(request_time)))

output:

  request_time
1   2020-12-31
2   2020-11-14

library(tidyr)
df %>% 
  separate(request_time, c("date", "time"), sep=" ", remove = FALSE)

        request_time       date     time
1 2020-12-31 13:05:00 2020-12-31 13:05:00
2 2020-11-14 02:04:01 2020-11-14 02:04:01

CodePudding user response：

The NULL part is not clear. If it is a string "NULL", the as.Date should return NA. Otherwise, NULL cannot exist as such. It may be a list column (which is not clear)

df$request_time <- as.Date(df$request_time)

-output

df$request_time
[1] "2020-12-31" NA           "2020-11-14"

data

df <- data.frame(request_time = c("2020-12-31 13:05:00", "NULL", "2020-11-14 02:04:01"))

CodePudding user response：

Just extract the date using str_extract:

library(stringr)
library(dplyr)
f %>%
  mutate(request_time = str_extract(request_time, "[0-9-] "))
  request_time
1   2020-12-31
2   2020-11-14

in base R:

f$request_time <- str_extract(f$request_time, "[0-9-] ")