[R] problem with as.Date

[Gabor Grothendieck]

On Sat, Oct 22, 2011 at 4:26 PM, Martin Spindler <Martin.Spindler at gmx.de> wrote:
> Dear all,
>
> I would like to convert the first column of a dataframe to a date (original
> format: year (4 digits) and month (last 2 digits))
>
>>str(dat_FF)
> 'data.frame':   1022 obs. of  4 variables:
>  $ date : int  192607 192608 192609 192610 192611 192612 192701 192702
> 192703 192704 ...
>  $ Rm.Rf: num  2.69 2.52 0 -3.06 2.42 2.66 0 4.29 0.51 0.57 ...
>  $ SMB  : num  -2.49 -1.25 -1.38 -0.2 -0.34 -0.07 -0.11 0.35 -1.87 0.44 ...
>  $ HML  : num  -2.91 4.25 0.22 0.71 -0.4 -0.11 4.92 3.17 -2.92 1.33 ...
>
> But
>
>>dat_FF$date <- as.Date(as.character(dat_FF$date), format="%Y%m")
>
> delievers NAs:
>
>>str(dat_FF)
> 'data.frame':   1022 obs. of  4 variables:
>  $ date : Date, format: NA NA ...
>
> I am very grateful for hints! Thanks in advance!
>

Here are two approaches:

# sample input
x <- c(192607, 192608, 192609, 192610, 192611, 192612, 192701, 192702)

# 1 - append day
as.Date(paste(x, 1), "%Y%m %d")

# 2 - yearmon
library(zoo)
as.Date(as.yearmon(format(x), "%Y%m"))

In the last one we use zoo's yearmon which can directly represent
year/month without a day and then we convert it to a date.  In fact
you might just want to leave it as a yearmon and use that directly
since it corresponds more closely to what you have:

> as.yearmon(format(x), "%Y%m")
[1] "Jul 1926" "Aug 1926" "Sep 1926" "Oct 1926" "Nov 1926" "Dec 1926" "Jan 1927"
[8] "Feb 1927"

See ?yearmon


-- 
Statistics & Software Consulting
GKX Group, GKX Associates Inc.
tel: 1-877-GKX-GROUP
email: ggrothendieck at gmail.com