Filter date column for certain months

1
votes

I would like to turn a monthly time series into a quarterly one, by only keeping 4 values from the year. I have a date column in a dataframe as an integer in this monthly format:

192707
192708
192709
....

I have converted it to date format:

library(zoo)
df$date<-as.Date(as.yearmon(as.character(df$date), format = '%Y%m'), frac=1)

Next I was thinking to keep only 4 months of the year (Jan, Apr, Jul, Oct) with something like this, but I am not sure how to exactly do this:

df<-subset(df, month(df$date)==1 | month(df$date)==4 | 
            month(df$date)==7 | month(df$date)==10)

Is there a way to extract the month from the date column?

2
rdatefilter
I'd recommend using a package. Specifically, see this vignette from the tibbletime package. - JasonAizkalns
You can get the month from a yearmon variable ym without converting it to Date using cycle(ym) . - G. Grothendieck
Thanks for the suggestion, tibbletime is very nice. cycle(ym) is also very nice!! - Niccola Tartaglia

2 Answers

4
votes

Yes, you are very close. Using the months function gives the full name of the month of the date, then using month.name we can get just the 4 months you want:

df <- df[months(df$date) %in% month.name[c(1,4,7,10)], ]
2
votes

Another option could be using month number i.e. 1, 4, 7, 10 is as:

df[month(df$date) %% 3 == 1,]

In this one can avoid comparison on name. Hence it should be quick.

Note: I must mention here that OP has used zoo package.