How to change multiple Date formats in same column

Question

What I've got so far is a dataframe column with dates in different character formats. A few appear in the %d.%m.%Y pattern, some in %m/%d/%Y :

data$initialDiagnose = as.character(data$initialDiagnose)
data$initialDiagnose[1:10]

[1] "14.01.2009" "9/22/2005"  "4/21/2010" "28.01.2010" "09.01.2009" "3/28/2005" "04.01.2005" "04.01.2005" "9/17/2010" "03.01.2010"

I want them as Date() in one format, but R refuses of course.
So I tried at first to change them by the separator:

data$initialDiagnose[grep('/', data$initialDiagnose)] = as.character.Date(data$initialDiagnose[grep('/', data$initialDiagnose)], format = '%m/%d/%Y')

Analogue to the '.' dates. But it didn't work.

How can I change them all to one format, that I can work with them?

Answer 1

a <- as.Date(data$initialDiagnose,format="%m/%d/%Y") # Produces NA when format is not "%m/%d/%Y"
b <- as.Date(data$initialDiagnose,format="%d.%m.%Y") # Produces NA when format is not "%d.%m.%Y"
a[is.na(a)] <- b[!is.na(b)] # Combine both while keeping their ranks
data$initialDiagnose <- a # Put it back in your dataframe
data$initialDiagnose
[1] "2009-01-14" "2005-09-22" "2010-04-21" "2010-01-28" "2009-01-09" "2005-03-28" "2005-01-04" "2005-01-04" "2010-09-17" "2010-01-03"

Additionnaly here's the preceding method adapted to a situation where you have three (or more) different formats:

data$initialDiagnose
[1] 14.01.2009 9/22/2005  12 Mar 97  4/21/2010  28.01.2010 09.01.2009 3/28/2005 
Levels: 09.01.2009 12 Mar 97 14.01.2009 28.01.2010 3/28/2005 4/21/2010 9/22/2005

multidate <- function(data, formats){
    a<-list()
    for(i in 1:length(formats)){
        a[[i]]<- as.Date(data,format=formats[i])
        a[[1]][!is.na(a[[i]])]<-a[[i]][!is.na(a[[i]])]
        }
    a[[1]]
    }

data$initialDiagnose <- multidate(data$initialDiagnose, 
                                  c("%m/%d/%Y","%d.%m.%Y","%d %b %y"))
data$initialDiagnose
[1] "2009-01-14" "2005-09-22" "1997-03-12" "2010-04-21" "2010-01-28" "2009-01-09" "2005-03-28"

Answer 2

I like lubridate for its ease of use:

library(lubridate) 

# note added ugly formats below
data <- data.frame(initialDiagnose = c("14.01.2009", "9/22/2005", 
        "4/21/2010", "28.01.2010", "09.01.2009", "3/28/2005", 
        "04.01.2005", "04.01.2005", "Created on 9/17/2010", "03 01 2010"))

mdy <- mdy(data$initialDiagnose) 
dmy <- dmy(data$initialDiagnose) 
mdy[is.na(mdy)] <- dmy[is.na(mdy)] # some dates are ambiguous, here we give 
data$initialDiagnose <- mdy        # mdy precedence over dmy
data
#   initialDiagnose
#       2009-01-14
#       2005-09-22
#       2010-04-21
#       2010-01-28
#       2009-09-01
#       2005-03-28
#       2005-04-01
#       2005-04-01
#       2010-09-17
#       2010-03-01