Filtering on a date using dplyr without changing the variable format

2
votes

I want use a list of years to filter a database by date

years<-c("2014")
yearsdata <- data.frame(animal=c("cow","pig"),
                        mydate=c(as.Date("2015-01-01"),
                        as.Date("2014-01-01")))
yearsdata %>% 
          mutate(mydate =format(mydate, "%Y") %>% 
          as.character()) %>%     
          filter(is.null(years) | mydate %in% years)

The above code works and lets me filter my dataset but it also formats the date column. Is there a way to get my filter results without having the format of the date column change in the finished subset dataframe?

2
rfilterdplyrmutated
If you don't want to format the date, then don't overwrite mydate with format(mydate). Just make a new column year = format(mydata, "%Y") and you can drop it at the end select(-year). (Or use lubridate like below.) - Gregor Thomas
Are your parentheses misplaced? You have a %>% inside your mutate()! - Frank

2 Answers

7
votes

If you're up for using the lubridate package you could do:

library("lubridate")

yearsdata %>%
  filter(is.null(years) | year(mydate) %in% years)

Which gives:

#   animal     mydate
# 1    pig 2014-01-01
5
votes

All these pipes gives me headache, I would just do

library(data.table)
setDT(yearsdata)[is.null(years) | year(mydate) %in% years]
#    animal     mydate
# 1:    pig 2014-01-01