Changing date formats in R

Dates come in all sorts of formats, but R generally prefers dates in the format YYYY-MM-DD. Here’s a quick method to convert date formats in R.

Let’s say we have a CSV file with dates in the format %d-%m-%Y (eg: 21/06/2021) and would like to convert these Australian-style dates into something more widely recognised internationally (ie: %Y-%m-%d). This isn’t as tricky as it seems.

Let’s begin by importing our data:

library(data.table)
df <- fread("data.csv")

If we use the head(df) function, we can see the structure of the data (which is very simple for this example):

         date         MB
1: 01/03/2021   9128.039
2: 02/03/2021 124144.419
3: 03/03/2021 135989.970
4: 04/03/2021   2925.900
5: 05/03/2021   3374.590
6: 06/03/2021   2186.940

The dates are all in %d-%m-%Y format. To change this, we firstly need to parse the data:

date <- strptime(as.character(df$date), "%d/%m/%Y")

Next, the dates are reformatted:

date <- format(date, "%Y-%m-%d")

The final step is to put the newly reformatted data into a new data frame. Since there is only one other column in this data frame, I’ll rename it as follows:

MB <- df$MB

Now I can put my columns into a new data frame:

df2 <- data.frame(date, MB)

Let’s check that with head(df2):

        date         MB
1 2021-03-01   9128.039
2 2021-03-02 124144.419
3 2021-03-03 135989.970
4 2021-03-04   2925.900
5 2021-03-05   3374.590
6 2021-03-06   2186.940

Voila! The date format in our dataset has been reformatted. Unfortunately R only recognises dates in the following two formats:

• %Y-%m-%d
• %Y/%m/%d

Anything else may lead to one of those annoying “character string is not in a standard unambiguous format” errrors whenever you start analysing your data. This why it’s important to correctly format your dates as soon as you open a data file in R.

Have Your Say

Δ