R as.POSIXct 转换为 as.Date

Question

I'm dealing with a data.frame object that holds in the first column a 'Date' field. Initially, I converted all the dates (which came in 'Date' as plain string) to POSIXct dates applying as.POSIXct() , but then I realized that as.Date was sufficient to my purpose. So I apply the as.Date() function to the POSIXct dates and I get a strange result: all the dates are scaled back one day ('2020-07-02 01:00:00' --> '2020-07-01'). Also tried with as.Date.POSIXct() with the same result.
Is there something I've missed in the conversion? Is this type of conversion legitim?
Thanks

Answer 1

Without knowing exactly what the data looks like, my best guess would have to do with the origin date used in the as.Date() function call when you convert from POSIXct. Hadley explains here the POSIXct and POSIXlt data types are really just double float values representing number of seconds since 1970-01-01. According to ?as.Date, if you're converting a value of class numeric, you should supply a value to the origin argument.

You could just reload your raw data, and play around with as.Date() and as.POSIXct() and see how the origin argument plays into an accurate conversion.