R help - Oct 2001 - creating chron object aggregates (e.g. sums by day)

What is the recommended/optimal way to perform aggregates on data frames
with chron objects?

Here is an example:
>raw.data
1 07/09/01   4000
2 07/09/01   2000
3 07/11/01   1000
4 07/13/01   800
5 07/13/01   700
6 07/16/01   600
7 07/17/01   500

I'm trying to construct a function that would first aggregate the data
(second column) by day (grouping by the first column) according to a
function (here "sum", but could be "max" or other)
>chronaggregate(raw.data, sum, "days")   #(used "days"
since 07/09/01 is
short for 07/09/01 00:00:00, but could be 07/09/01 00:12:34)
1 07/09/01   6000      << sum of data values for day 07/09/01 from
raw.data
2 07/11/01   1000
3 07/13/01   1500      << sum of data values for day 07/13/01 from
raw.data
4 07/16/01   600
5 07/17/01   500

and insert 0 values for days without data:

1 07/09/01   6000
2 07/10/01   0              << inserted record
3 07/11/01   1000
4 07/12/01   0              << inserted record
5 07/13/01   1500
6 07/14/01   0              << inserted record
7 07/15/01   0              << inserted record
8 07/16/01   600
9 07/17/01   500

Is there a simple way to do this?

Thanks,

--
-Olivier

--
Olivier Collignon
Loudcloud, Inc.
olivier at loudcloud.com



-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

hallo,

the following code isn't very pretty, but it works. The great problem is
the format of the date.

#read your data
raw.data <- read.table('yourdatafile')
#collect the values of a single day (defined by the first 8 characters
of your date-code)
sum.raw.data <- tapply(raw.data[,2],substring(raw.data[,1],1,8),sum)
#creating the date range of your final variable 
dr <- as.character(seq(ISOdate(2001,7,1,0), ISOdate(2001,7,31,0),
"days"))
#Change the format of them to the same of your input data
sdr <-
paste(substring(dr,6,7),substr(dr,9,10),substr(dr,3,4),sep='/')
#creating a vector with the correct length
arr.data <- rep(0,length(sdr))
#set the names of the single elements
names(arr.data) <- sdr
#now you can access the elements direct with their name
arr.data[names(sum.raw.data)] <- sum.raw.data
#show the result
arr.data


an other (similary) way is, to change the date format of your data

raw.data <- read.table('test.txt')
sum.raw.data <- tapply(raw.data[,2],substring(raw.data[,1],1,8),sum)
new.date <- strptime(names(sum.raw.data),format="%m/%d/%y")
#convert
date format

se <- seq(ISOdate(2001,7,1,0), ISOdate(2001,7,31,0), "days")
arr.data <- rep(0,length(se))
names(arr.data) <- substr(as.character(se),1,10)
arr.data[as.character(new.date)] <- sum.raw.data


it's this simple enough?

gruess

joerg ameder


Olivier Collignon wrote:
> 
> What is the recommended/optimal way to perform aggregates on data frames
> with chron objects?
> 
> Here is an example:
> 
> >raw.data
> 1 07/09/01   4000
> 2 07/09/01   2000
> 3 07/11/01   1000
> 4 07/13/01   800
> 5 07/13/01   700
> 6 07/16/01   600
> 7 07/17/01   500
> 
> I'm trying to construct a function that would first aggregate the data
> (second column) by day (grouping by the first column) according to a
> function (here "sum", but could be "max" or other)
> 
> >chronaggregate(raw.data, sum, "days")   #(used
"days" since 07/09/01 is
> short for 07/09/01 00:00:00, but could be 07/09/01 00:12:34)
> 1 07/09/01   6000      << sum of data values for day 07/09/01 from
> raw.data
> 2 07/11/01   1000
> 3 07/13/01   1500      << sum of data values for day 07/13/01 from
> raw.data
> 4 07/16/01   600
> 5 07/17/01   500
> 
> and insert 0 values for days without data:
> 
> 1 07/09/01   6000
> 2 07/10/01   0              << inserted record
> 3 07/11/01   1000
> 4 07/12/01   0              << inserted record
> 5 07/13/01   1500
> 6 07/14/01   0              << inserted record
> 7 07/15/01   0              << inserted record
> 8 07/16/01   600
> 9 07/17/01   500
> 
> Is there a simple way to do this?
> 
> Thanks,
> 
> --
> -Olivier
> 
> --
> Olivier Collignon
> Loudcloud, Inc.
> olivier at loudcloud.com
> 
>
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
> r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
> Send "info", "help", or "[un]subscribe"
> (in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
>
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
-- 
    Joerg Maeder             IACETH              INSTITUTE
   PhD Student                              FOR ATMOSPHERIC 
  Phone: +41 1 633 36 25                 AND CLIMATE SCIENCE
 Fax: +41 1 633 10 58                  ETH Z?RICH Switzerland
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

Hello,

I have an alternative code to resolve your problem. Not as simple and 
elegant as I would like, but quite easy to understand and maintain. Here it 
is :

# reading your data, preventing dates to be converted as factor
raw.data <- read.table("your data file", as.is=TRUE)
# aggregating your observations by day, using a numerical conversion of the 
dates
agg.data.1 <- tapply(raw.data[[2]], as.numeric(dates(raw.data[[1]])), sum)
agg.data.1
# moving from the results' table from tapply back to a data frame with 
named rows
agg.data.2 <- data.frame(sumval = agg.data.1[1:length(agg.data.1)])
agg.data.2
# creating the (numerical) date range of your observations
date.stamp <- 
data.frame(date=min(as.numeric(row.names(agg.data.2))):max(as.numeric(row.names(agg.data.2))))
date.stamp
# merging the aggregated data and the dates stamps
agg.data.3 <- merge(date.stamp, agg.data.2, by.x="date",
by.y="row.names",
all.x=TRUE)
# replacing the NA's where there are no value by zero
agg.data.3$sumval[is.na(agg.data.3$sumval)] <- 0
# converting back the numericals to dates
agg.data.3$date <- dates(agg.data.3$date)
agg.data.3

Hope it helped a bit !

Yves.

====================================================================        
YVES BROSTAUX - Ing?nieur agronome Orientation Eaux & For?ts

         Assistant - Unit? de Statistique et Informatique
         Gembloux Agricultural University
         8, avenue de la Facult? B-5030 Gembloux (Belgium)
         T?l:      +32 (0)81 62 24 69
         E-mail :  brostaux.y at fsagx.ac.be
====================================================================
At 04:01 31/10/01, you wrote:
>Olivier Collignon wrote:
> >
> > What is the recommended/optimal way to perform aggregates on data
frames
> > with chron objects?
> >
> > Here is an example:
> >
> > >raw.data
> > 1 07/09/01   4000
> > 2 07/09/01   2000
> > 3 07/11/01   1000
> > 4 07/13/01   800
> > 5 07/13/01   700
> > 6 07/16/01   600
> > 7 07/17/01   500
> >
> > I'm trying to construct a function that would first aggregate the
data
> > (second column) by day (grouping by the first column) according to a
> > function (here "sum", but could be "max" or other)
> >
> > >chronaggregate(raw.data, sum, "days")   #(used
"days" since 07/09/01 is
> > short for 07/09/01 00:00:00, but could be 07/09/01 00:12:34)
> > 1 07/09/01   6000      << sum of data values for day 07/09/01
from
> > raw.data
> > 2 07/11/01   1000
> > 3 07/13/01   1500      << sum of data values for day 07/13/01
from
> > raw.data
> > 4 07/16/01   600
> > 5 07/17/01   500
> >
> > and insert 0 values for days without data:
> >
> > 1 07/09/01   6000
> > 2 07/10/01   0              << inserted record
> > 3 07/11/01   1000
> > 4 07/12/01   0              << inserted record
> > 5 07/13/01   1500
> > 6 07/14/01   0              << inserted record
> > 7 07/15/01   0              << inserted record
> > 8 07/16/01   600
> > 9 07/17/01   500
> >
> > Is there a simple way to do this?
> >
> > Thanks,
> >
> > --
> > -Olivier
> >
> > --
> > Olivier Collignon
> > Loudcloud, Inc.
> > olivier at loudcloud.com
> >
> > 
>
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
> > r-help mailing list -- Read
http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
> > Send "info", "help", or "[un]subscribe"
> > (in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
> > 
>
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._