Modify the date in a data frame in R - r

Question

Recently I stumble over a problem. Unfortunately my variable for the date has not been recorded uniformly.
I got a similar data frame like the one shown below
Variable1 <- c(10,20,30,40,50)
Variable2 <- c("a", "b", "c", "d", "d")
Date <- c("today 10:45", "yesterday 3:10", "28 october 2018 5:32", "28 october 2018 8:32", "27 october 2018 5:32")
df <- data.frame(Variable1, Variable2, Date)
df
For my use I need to extract only the date of it. Therefore, I would like to create a new variable based on "Date".
The Date variable should only contain the date. The hour is irrelevant for my purpose and can be ignored.
My goal is to get the following data frame:
Variable1 <- c(10,20,30,40,50)
Variable2 <- c("a", "b", "c", "d", "d")
Date <- c("31 october 2018", "30 october 2018", "28 october 2018", "28 october 2018", "27 october 2018")
df2 <- data.frame(Variable1, Variable2, Date)
df2
Preferably the values for Date should also be in the correct format (date).
Thank you already in advance.

Answer 1

df$NewDate[grepl("today",df$Date)]<-Sys.Date() # Convert today to date
df$NewDate[grepl("yesterday",df$Date)]<-Sys.Date()-1 # covert yesterday to date
df$NewDate[is.na(df$NewDate)]<-df$Date[is.na(df$NewDate)] %>% as.Date(format="%d %b %Y") # Convert explicit dates to date format
class(df$NewDate)<-"Date" # Convert column to Date class
df
Variable1 Variable2 Date NewDate
1 10 a today 10:45 2018-10-31
2 20 b yesterday 3:10 2018-10-30
3 30 c 28 october 2018 5:32 2018-10-28
4 40 d 28 october 2018 8:32 2018-10-28
5 50 d 27 october 2018 5:32 2018-10-27

Answer 2

tolower( # not strictly necessary, but for consistency
gsub("yesterday", format(Sys.Date()-1, "%d %B %Y"), # convert *day to dates
gsub("today", format(Sys.Date(), "%d %B %Y"),
gsub("\\s*[0-9:]*$", "", # remove the times
c("today 10:45", "yesterday 3:10", "28 october 2018 5:32", "28 october 2018 8:32", "27 october 2018 5:32")))))
# [1] "31 october 2018" "30 october 2018" "28 october 2018" "28 october 2018" "27 october 2018"

Answer 3

Another solution, using indices.
Date <- c("today 10:45", "yesterday 3:10", "28 october 2018 5:32", "28 october 2018 8:32", "27 october 2018 5:32")
Date <- sub("today", Sys.Date(), Date)
Date <- sub("yesterday", Sys.Date() - 1, Date)
i <- grep("[[:alpha:]]", Date)
Date[i] <- format(as.POSIXct(Date[i], format = "%d %B %Y %H:%M"), format = "%d %B %Y")
Date[-i] <- format(as.POSIXct(Date[-i]), format = "%d %B %Y")
Date
#[1] "31 October 2018" "30 October 2018" "28 October 2018"
#[4] "28 October 2018" "27 October 2018"
Then I noticed the solution by user r2evans, that converts everything to lowercase. So, if necessary, end with
Date <- tolower(Date)