R help - Jun 2004 - Aggregating on Water Year Rather Than Calendar Year

The US water year extends from 01 October yyyy-1 through 30 September yyyy
and is referenced by the year starting on the included 01 January yyyy.  
I'd like to be able to find the annual means for the water year.  To do so 
I've taken the input date-time, which is in the usual format 

"1991-10-07 10:35:00"

changed it by:

w$d<-as.POSIXct(w$date.time)

Now I can add an offset of 92 days 

w$w.year<-as.POSIXct(w$d+7948800)

and have the years correspond to the "water year."

Now I wish to obtain some means by something like:

waterT<-aggregate(w$value[w$param.name=="Temperature"],
list(w$w.year[w$param.name=="Temperature"]),mean)

Except that I need to work on the year and being a neophyte in date 
arithmetic I'm not finding a working method.

TIA

Clint

-- 
Clint Bowman			INTERNET:	clint at ecy.wa.gov
Air Quality Modeler		INTERNET:	clint at math.utah.edu
Department of Ecology		VOICE:		(360) 407-6815
PO Box 47600			FAX:		(360) 407-7534
Olympia, WA 98504-7600

It's not clear where your problem is.

Did w$w.year come out wrong?
Or did the aggregate() function fail?

For getting the water year, I would do it differently:
  (this is for a U.S. English locale, and minimally tested)

    tmp <- as.POSIXlt(w$d)$year+1900
    w$w.year <- ifelse( format(w$d,'%b') %in%
c('Oct','Nov','Dec'), tmp+1, tmp)

Then you can do things like
   table(w$w.year)
   table(w$w.year,w$param.name)
to help find out if w.year came out like it should.

You'll have to provide error messages or something if the problem is 
with using aggregate().

At 11:21 AM -0700 6/16/04, Clint Bowman wrote:
>The US water year extends from 01 October yyyy-1 through 30 September yyyy
>and is referenced by the year starting on the included 01 January yyyy. 
>I'd like to be able to find the annual means for the water year.  To do
so
>I've taken the input date-time, which is in the usual format
>
>"1991-10-07 10:35:00"
>
>changed it by:
>
>w$d<-as.POSIXct(w$date.time)
>
>Now I can add an offset of 92 days
>
>w$w.year<-as.POSIXct(w$d+7948800)
Try   w$w.year <- w$d+7948800
>
>and have the years correspond to the "water year."
>
>Now I wish to obtain some means by something like:
>
>waterT<-aggregate(w$value[w$param.name=="Temperature"],
>list(w$w.year[w$param.name=="Temperature"]),mean)
>
>Except that I need to work on the year and being a neophyte in date
>arithmetic I'm not finding a working method.
>
>TIA
>
>Clint
>
>--
>Clint Bowman			INTERNET:	clint at ecy.wa.gov
>Air Quality Modeler		INTERNET:	clint at math.utah.edu
>Department of Ecology		VOICE:		(360) 407-6815
>PO Box 47600			FAX:		(360) 407-7534
>Olympia, WA 98504-7600
>
-Don

-- 
--------------------------------------
Don MacQueen
Environmental Protection Department
Lawrence Livermore National Laboratory
Livermore, CA, USA

Try this to get the US water year:

require(chron)
with(month.day.year(unclass(as.Date(w$date.time))), year+(month>9))

Date:   Wed, 16 Jun 2004 11:21:56 -0700 (PDT) 
From:   Clint Bowman <clint at ecy.wa.gov>
To:   <r-help at stat.math.ethz.ch> 
Subject:   [R] Aggregating on Water Year Rather Than Calendar Year 

 
The US water year extends from 01 October yyyy-1 through 30 September yyyy
and is referenced by the year starting on the included 01 January yyyy. 
I'd like to be able to find the annual means for the water year. To do so 
I've taken the input date-time, which is in the usual format 

"1991-10-07 10:35:00"

changed it by:

w$d<-as.POSIXct(w$date.time)

Now I can add an offset of 92 days 

w$w.year<-as.POSIXct(w$d+7948800)

and have the years correspond to the "water year."

Now I wish to obtain some means by something like:

waterT<-aggregate(w$value[w$param.name=="Temperature"],
list(w$w.year[w$param.name=="Temperature"]),mean)

Except that I need to work on the year and being a neophyte in date 
arithmetic I'm not finding a working method.

TIA

And here is a slightly shorter variant:

   with(as.POSIXlt(w$date.time), 1900+year+(mon>8) )

---
Gabor Grothendieck <ggrothendieck at myway.com>:

Try this to get the US water year:

require(chron)
with(month.day.year(unclass(as.Date(w$date.time))), year+(month>9))

Date: Wed, 16 Jun 2004 11:21:56 -0700 (PDT) 
From: Clint Bowman <clint at ecy.wa.gov>
To: <r-help at stat.math.ethz.ch> 
Subject: [R] Aggregating on Water Year Rather Than Calendar Year 


The US water year extends from 01 October yyyy-1 through 30 September yyyy
and is referenced by the year starting on the included 01 January yyyy. 
I'd like to be able to find the annual means for the water year. To do so 
I've taken the input date-time, which is in the usual format 

"1991-10-07 10:35:00"

changed it by:

w$d<-as.POSIXct(w$date.time)

Now I can add an offset of 92 days 

w$w.year<-as.POSIXct(w$d+7948800)

and have the years correspond to the "water year."

Now I wish to obtain some means by something like:

waterT<-aggregate(w$value[w$param.name=="Temperature"],
list(w$w.year[w$param.name=="Temperature"]),mean)

Except that I need to work on the year and being a neophyte in date 
arithmetic I'm not finding a working method.

TIA