R delete duplicate elements in character vector, not duplicate lines

Question

I ran into a brick wall with this problem.

I have a data (date) frame with some document identifiers and dates stored in a character vector:

Doc Dates 1 12345 c("06/01/2000","08/09/2002") 2 23456 c("07/01/2000", 09/08/2003", "07/01/2000") 3 34567 c("09/06/2004", "09/06/2004", "12/30/2006") 4 45678 c("06/01/2000","08/09/2002") 

I am trying to remove duplicate elements in dates to get this result:

  Doc Dates 1 12345 c("06/01/2000","08/09/2002") 2 23456 c("07/01/2000", 09/08/2003") 3 34567 c("09/06/2004", "12/30/2006") 4 45678 c("06/01/2000","08/09/2002") 

I tried:

 R>unique(dates$dates) 

but removes duplicate rows by dates:

  Doc Dates 1 12345 c("06/01/2000","08/09/2002") 2 23456 c("07/01/2000", 09/08/2003") 3 34567 c("09/06/2004", "12/30/2006") 

Any help on how to remove only duplicate items in dates, and not remove duplicate rows by dates?

** Updated with data

 # Match some text string (dates) from some text: df1$dates <- as.character(strapply(df1[[2]], "((\\D\\d{1,2}(/|-)\\d{1,2}(/|-)\\d{2,4})| ([^/]\\d{1,2}(/|-)\\d{2,4})|((JAN|FEB|MAR|APR|MAY|JUN|JUL|AUG|SEP|OCT|NOV){1}[\\s|-]{0,2}\\d{1,4}(\\D[\\s|-]{0,}\\d{2,4}){0,}))")) # Drop first 2 columns from dataframe df2<-df1[ -c(1,2)] # List data >df2 872 7/23/2007 873 c(" 11/4/2007", " 11/4/2007") 874 c(" 4/2/2008", " 8/2/2007") 880 11/14/2006 > class(df2) [1] "data.frame" > class(df2$dates) [1] "character" > dput(df2) structure(list(dates = c("NULL", "NULL", " 7/23/2007", "c(\" 11/4/2007\", \" 11/4/2007\")", "c(\" 4/2/2008\", \" 8/2/2007\")", "NULL", "NULL", "NULL", "NULL", "NULL", " 11/14/2006")), .Names = "dates", class = "data.frame", row.names = 870:880) 

So my problem is how to get rid of duplicate dates on line 873?

Answer 1

Try the following:

 within(dates, Dates <- lapply(Dates, unique)) 

Answer 2

I solved the problem associated with removing duplicate values ​​from a character vector - wrap lapply (binding (), unique):

 df1$date <- as.character(lapply((strapply(df1[[2]], "((\\D\\d{1,2}(/|-)\\d{1,2}(/|- )\\d{2,4})|(\\s\\d{1,2}(/|-)\\d{2,4})|((JAN|FEB|MAR|APR|MAY|JUN|JUL|AUG|SEP|OCT|NOV){1}[\\s|-]{0,2}\\d{1,4}(\\D[\\s|-]{0,}\\d{2,4}){0,}))")),unique)) 

Thank you for your help.

Answer 3

I would gsub out c( ) in the dates, and then for each line I would call unique on strsplit using ,

UNTESTED, but it could be something like: sapply(dates$dates, function(x){ new.x=gsub("c(|)","",x) new.x=strsplit(new.x, ",") unique(new.x) })

Answer 4

Perhaps you are looking for something like this.

  df Doc Dates 1 12345 c("06/01/2000","08/09/2002") 2 23456 c("07/01/2000", "09/08/2003", "07/01/2000") 3 34567 c("09/06/2004", "09/06/2004", "12/30/2006") 4 45678 c("06/01/2000","08/09/2002") Eval and Parse x <- t(sapply(df[,"Dates"],function(x){unique(eval(parse(text = x)))})) df$Dates <- paste(x[,1],x[,2],sep=",") df Doc Dates 1 12345 06/01/2000,08/09/2002 2 23456 07/01/2000,09/08/2003 3 34567 09/06/2004,12/30/2006 4 45678 06/01/2000,08/09/2002 Same can be achieved using Regex: paste(unique(unlist(strsplit(gsub("c\\(|\\)","",'c("24/07/2012","22/01/2012","24/07/2012")'),","))),sep = "") [1] "\"24/07/2012\"" "\"22/01/2012\"" Haven't tried on data but works