s-news
[Top] [All Lists]

Re: Merging Time Series

To: "Horace Tso" <Horace_Tso@pgn.com>, <s-news@lists.biostat.wustl.edu>
Subject: Re: Merging Time Series
From: "Pope, Alun" <Alun.Pope@apra.gov.au>
Date: Fri, 20 Jun 2003 14:53:53 +1000
Thread-index: AcM2yuYMaU4JlXVUTi6cG4Kh1Z2klgAG+3WA
Thread-topic: [S] Merging Time Series
I have also had this problem recently and found a solution.  I don't have time 
to re-edit it so in case it's helpful, here it is in its more or less original 
form.
Generally wouldn't it be good to have some simple ways of handling timeSeries 
and timeDate objects?


#the following 4 time series have different lengths and different frequencies 
(some quarterly some annual)
Banks.Med.pct.ts<-timeSeries(Banks.Med.pct,BanksTime,units="Percent")
Banks.Mean.pct.ts<-timeSeries(Panama.10JuneB$Total.Exposure,BanksTime,units="Percent")
CU.Med.ts<-timeSeries(CU.Med.pct,BSTime,units="Percent")
BS.ts<-timeSeries(BS.Med.pct,BSTime,units="Percent")

#the following does exactly what it ought to do, but I cannot see how to 
extract the multiple time series from the 4:
plot(Banks.Med.pct.ts,Banks.Mean.pct.ts,CU.Med.ts,BS.ts)

#this below enables use of ts.union but loses the timeSeries features:

Banks.cts<-cts(cbind(Banks.Med.pct,Panama.10JuneB$Total.Exposure),dates("01/28/1990"),"months",names=c("Median","Average"))

CUBS.cts<-cts(cbind(CU.Med.pct,BS.Med.pct),dates("03/28/1993"),"quarters",names=c("Credit
 Unions","Building Societies"))


#the fllowing objects have the correct behaviour
ADIs.all<-ts.union(Banks.cts,CUBS.cts)

Hope this helps.

Alun

____________________
Dr Alun Pope

Statistics Consultant
Australian Prudential Regulation Authority
400 George Street, Sydney NSW 2000, Australia
GPO Box 9836, Sydney NSW 2001, Australia

Tel: (02) 9210 3232
Fax: (02) 9210 3021
E-mail: alun.pope@apra.gov.au

                -----Original Message-----
                From: Horace Tso [mailto:Horace_Tso@pgn.com] 
                Sent: Friday, 20 June 2003 11:24 AM
                To: s-news@lists.biostat.wustl.edu
                Subject: [S] Merging Time Series

                Hi folks,
                 
                Using 6.1 under Win2K, I want to do a Union of two Time Series. 
Sounds simple? I've tried every trick out there and nothing works. SPlus 
support suggests using merge(), but that doesn't work. Each series has 
identical column names, Y1, Y2, Y3. There are some overlapping dates in both. 
Here they are, 
                 
                > t1
                Position    Y1    Y2      Y3
                02/12/1997 7.00  6.50 7.50 
                02/26/1997 6.50  6.25 6.75 
                02/27/1997 6.50  6.25 6.75
                02/28/1997 6.50  6.25 6.75
                03/01/1997 6.50  6.25 6.75
                03/03/1997 6.50  6.25 6.75
                03/04/1997 9.21  7.25 10.00
                03/05/1997 10.00 9.75  10.25
                03/06/1997 8.75  8.50  9.00
                 
                >t2
                Position    Y1    Y2      Y3
                02/12/1997 6.00  8.50 7.50 
                02/26/1997 6.15  6.00 6.75 
                02/27/1997 4.50  4.25 4.75
                02/28/1997 3.50  3.25 3.25
                03/01/1997 6.50  6.25 6.75
                03/04/1997 6.50  6.25 6.75
                03/05/1997 1.21  2.25 1.00
                03/08/1997 1.00  0.75  0.25
                03/09/1997 2.25  2.50  2.00
                 
                I want to create the following,
                 
                Position         Y1    Y2      Y3    Y11    Y21    Y31
                02/12/1997  6.00  8.50 7.50  7.00   6.50   7.50
                02/26/1997  6.15  6.00 6.75  6.50   6.25   6.75
                02/27/1997  4.50  4.25 4.75  6.50   6.25    6.75
                02/28/1997  3.50  3.25 3.25  6.50   6.25    6.75
                03/01/1997  6.50  6.25 6.75  6.50   6.25    6.75
                03/03/1997  NA     NA   NA     6.50   6.25    6.75
                03/04/1997  6.50  6.25 6.75   9.21  7.25   10.00
                03/05/1997  1.21  2.25 1.00  10.00 9.75   10.25
                03/06/1997  NA     NA   NA     8.75    8.50   9.00
                03/08/1997  1.00  0.75   0.25 NA     NA      NA
                03/09/1997  2.25  2.50   2.00 NA     NA      NA
                 
                Thanks a lot.
                 
                Horace W. Tso
                Senior Quantitative Analyst
                Portland General Electric
                Portland, Oregon
                503-464-8430 (D)
                 
                 


IMPORTANT NOTICE:

This e-mail is intended solely for the person or organisation to whom it is
addressed, and may contain secret, confidential or legally privileged
information. 

If you have received this e-mail in error or are aware that you are not
authorised to have it, you MUST NOT use or copy it, or disclose its contents
to any person. If you do any of these things, you may be sued or prosecuted.

If you have received this e-mail in error, please contact the sender
immediately. 

<Prev in Thread] Current Thread [Next in Thread>