CODEWITHSUNDEEP
rdatedatetime
James B
James B

Reputation: 51

Parsing non-conventional dates in R

I hope you guys are having a good night.

I have a dataset that looks something like this:

data1
       [date]    [fuel_price] 
[1,]  201004         3.56           
[2,]  201105         3.32          
[3,]  201106         3.45           
[4,]  201212        3.94        
[4,]  201301         3.94

The dates are in the format of: YYYYMM, and I'm trying to have R interpret it that way. Here is my code so far:

DateNum = as.numeric(as.POSIXct(data1$date, format="%Y%m"))

PriceFun = approxfun(thedata1$fuel_price ~ DateNum)

PriceFromDate = function(x) {
  round(PriceFun(x), 3) }

Any reason why it isn't working for me?

Thank you guys in advanced.

Upvotes: 2

Views: 45

Answers (1)

Konrad
Konrad

Reputation: 18585

It's rather risky approach but you could consider making use of the anytime package:

  x <- c(201004,           
         201105,          
         201106,           
         2012012,        
         201301)  

  anytime::anytime(x, asUTC = TRUE)

this would give you:

[1] "2010-04-01 01:00:00 BST" "2011-05-01 01:00:00 BST" "2011-06-01 01:00:00 BST"
[4] "2012-12-01 00:00:00 GMT" "2013-01-01 00:00:00 GMT"

The points discussed in comments are valid, you may first prefer to arrive at unambiguous strings making conscious assumptions about the date format; like assuming first day of a month for each date and the undertaking your conversions; the provided solutions uses your initial string 2012012 with typo to demonstrate that anytime would manage this string as well.

Upvotes: 1

Related Questions