How to reverse a mutate command in R?

Question

I am doing the Bellabeat case study from the google analytics course. I am doing it in R and noticed that the date format was in character format rather than "Date" so I tried to use the mutate() function to convert it to a date format.

The dataframe is called sleep

The code looks as:

sleep <- sleep %>%
    mutate(date = ymd(Date))

But instead of converting it to a date format it went from this:

         Id       Date TotalSleepRecords TotalMinutesAsleep TotalTimeInBed
1 1503960366 4/12/2016                  1                327            346
2 1503960366 4/13/2016                  2                384            407
3 1503960366 4/15/2016                  1                412            442
4 1503960366 4/16/2016                  2                340            367
5 1503960366 4/17/2016                  1                700            712
6 1503960366 4/19/2016      

to,

> head(sleep)
          Id Date TotalSleepRecords TotalMinutesAsleep TotalTimeInBed date
1 1503960366 <NA>                 1                327            346 <NA>
2 1503960366 <NA>                 2                384            407 <NA>
3 1503960366 <NA>                 1                412            442 <NA>
4 1503960366 <NA>                 2                340            367 <NA>
5 1503960366 <NA>                 1                700            712 <NA>
6 1503960366 <NA>                 1                304            320 <NA>

What exactly did I do wrong and how can I revert it?

My purpose with the dataframes I work with is to compare how exercise impacts sleep. I am not sure if I actually need to convert the date format, but many others who did the case study did it, so I figured that it might impact the analysis if I do not do it. There are several other dataframes in the package I work with. I figure I need to merge dataframes to compare columns.

Answer 1

As mentioned in the replies, since your dates are in "mm/dd/yyyy" format you should use lubridate::mdy() to convert them to dates.

As a general workflow tip, it's a good idea to check what the results of your operation would be before overwriting the dataframe. For example if I have

library(dplyr)
library(lubridate)

sleep <- tibble(id = c(1, 2), date = c("4/12/2016", "4/13/2016"))

then in the console I can try out the operation first:

sleep %>%
   mutate(date = mdy(date))

#>      id date      
#>   <dbl> <date>    
#> 1     1 2016-04-12
#> 2     2 2016-04-13

Now that I'm satisfied my mutation has done what I want it to, I can save the results. I tend to save the results to a new dataframe:

sleep_clean <- sleep %>%
   mutate(date = mdy(date))

so that I can still access the original sleep dataframe after (thus 'reverting' to the previous version).