R help - Jun 2010 - big big problem

Dear list,
I'll try to be more clear in explaining my problem. I have a data frame like
this called X:



                        CLUSTER        YEAR      variable         value1        
value2




M1                     2005         EC01           NA              NA

M1                     2006         EC01            2                5

M1                     2007         EC01            4                5

M2                     2005          EC01          NA              NA

M2                     2006          EC01           5                7
M2                     2007          EC01           8                 7

M1                     2005          EC02           NA               NA

M1                     2006          EC02            3                 8

M1                    2007            EC02           1                 8

M2                    2005            EC02           NA               NA

M2                    2006            EC02           9                 10

M2                    2007            EC02            6                 10




I'm trying to build the time series of the variables (EC01,EC02) by applying
the following formulas in a recursive way(by starting from value2 at time 2007).
The formulas needed to get the time series are

series(2006)=value2(2007)-value1(2007)

series(2005)=value2(2006)-value1(2006)
In terms of number I would have for example for variable EC01:
series(2006)=5-4=1
series(2005)=1-2=-1
 In addition, this calculations should be down grouping by variable e cluster.So
the result should be
CLUSTER        YEAR      variable        series

M1                     2005         EC01             -1                


M1                     2006         EC01             1


M1                     2007         EC01              5


M2                     2005          EC01             -6



M2                     2006          EC01             -1

M2                     2007          EC01              7



M1                     2005          EC02              4

M1                     2006          EC02               7


M1                    2007            EC02              8


M2                    2005            EC02             -5


M2                    2006            EC02            4


M2                    2007            EC02             10
I applied the following formula which gives what I want but not at all:
series=ddply(X,.(variable,CLUSTER),transform,series=rev(value2-cumsum(rev(value1[-1]))))))
The first problems is that I get the following warning message:
longer object length is not a multiple of shorter object length
Secondly If I try to add new years(My aim was to do a check in order to see if
this procedure works for all type of data frame)
R does the right calculations but the order of the results referred to the
variable called series is not right.



Thanks for your attention!!!

















	[[alternative HTML version deleted]]

It seems to me you're not aware of what that code is doing. You can
easily add the 5 by :

ddply(X,.(variable,CLUSTER),transform,series=c(rev(value2[-1]-cumsum(rev(value1[-1]))),rev(value2)[1]))

Try to figure out what the code is doing exactly for next time, you'll
definitely benefit from it.
Cheers
Joris


On Thu, Jun 17, 2010 at 4:29 PM, n.vialma at libero.it <n.vialma at
libero.it> wrote:
> Well English is not my mother tongue too and for me its very difficult to
> explain the procedure also because the procedure is not easy.
> Anyway, with the command that you gave me I solved the problem, by putting
0
> this code works for all type of data frame
>
series=ddply(X,.(variable,CLUSTER),transform,series=rev(value2-c(0,cumsum(rev(value1[-1]))))).
> Concerning the formulas I provide for the construction of the time series,
> they are right but? another way to get the same results is to make
> series=rev(value2-cumsum(rev(value1[-1]))))),so by using the cumsum
> function.
> To be honest my data are different from what I put in my example and I
didnt
> check? If the code would have worked for the data frame given as example.
> I have a question for you, suppose that the result that you get with this
> formula
>
>
series=ddply(X,.(variable,CLUSTER),transform,series=rev(value2-c(0,cumsum(rev(value1[-1])))))
> are right. What I get (in my real data set) is not this:
> ??? --
>
> TABLE1
>
> ??????????????? CLUSTER??????? YEAR????? variable??????? series
>
> ??????????????? M1???????????????????? 2005???????? EC01
> -1
>
> ??????????????? M1???????????????????? 2006???????? EC01???????????? 1
>
> ??????????????? M1???????????????????? 2007???????? EC01????????????? 5
>
> ??????????????? M2???????????????????? 2005????????? EC01???????????? -6
>
> ??????????????? M2???????????????????? 2006????????? EC01???????????? -1
>
> ??????????????? M2???????????????????? 2007????????? EC01????????????? 7
>
> ??????????????? M1???????????????????? 2005????????? EC02????????????? 4
>
> ??????????????? M1???????????????????? 2006????????? EC02?????????????? 7
>
> ??????????????? M1??????????????????? 2007??????????? EC02????????????? 8
>
> ??????????????? M2??????????????????? 2005??????????? EC02???????????? -5
>
> ??????????????? M2??????????????????? 2006??????????? EC02??????????? 4
>
> ??????????????? M2??????????????????? 2007??????????? EC02???????????? 10
>
>
> but this: TABLE2
>
> ???????????? CLUSTER??????? YEAR????? variable??????? series
>
> ??????????????? M1???????????????????? 2005???????? EC01
> -1
>
> ??????????????? M1???????????????????? 2006???????? EC01???????????? 1
>
> ??????????????? M1???????????????????? 2007???????? EC01????????????? NA
>
> ??????????????? M2???????????????????? 2005????????? EC01???????????? -6
>
> ??????????????? M2???????????????????? 2006????????? EC01???????????? -1
>
> ??????????????? M2???????????????????? 2007????????? EC01????????????? NA
>
> ??????????????? M1???????????????????? 2005????????? EC02????????????? 4
>
> ??????????????? M1???????????????????? 2006????????? EC02?????????????? 7
>
> ??????????????? M1??????????????????? 2007??????????? EC02????????????? NA
>
> ??????????????? M2??????????????????? 2005??????????? EC02???????????? -5
>
> ??????????????? M2??????????????????? 2006??????????? EC02??????????? 4
>
> ??????????????? M2??????????????????? 2007??????????? EC02???????????? NA
>
> So I dont have in my final data frame the value of the series at time 2007
> which should be? what I called value1 at time 2007 in the "original
data
> frame"? (as shown in table 1) which is this one (that I gave as
example):
>
> CLUSTER??????? YEAR????? variable ? ? ??? value1 ???????? value2
>
> M1???????????????????? 2005???????? EC01?????????? NA ? ? ? ? ? ?? NA
>
>
> M1???????????????????? 2006???????? EC01??????????? 2??????????????? 5
>
> M1???????????????????? 2007???????? EC01??????????? 4??????????????? 5
>
> M2???????????????????? 2005????????? EC01????????? NA????????????? NA
>
> M2???????????????????? 2006????????? EC01?????????? 5 ? ? ? ? ? ???? 7
>
>
> M2???????????????????? 2007????????? EC01?????????? 8???????????????? 7
>
> M1???????????????????? 2005????????? EC02 ????????? NA?????????????? NA
>
> M1???????????????????? 2006????????? EC02??????????? 3???????????????? 8
>
> M1??????????????????? 2007??????????? EC02?????????? 1???????????????? 8
>
> M2??????????????????? 2005??????????? EC02?????????? NA?????????????? NA
>
> M2??????????????????? 2006??????????? EC02?????????? 9???????????????? 10
>
> M2??????????????????? 2007??????????? EC02??????????? 6???????????????? 10
>
>
>
> I put 1 instead of 0 in the code
>
>
series=ddply(X,.(variable,CLUSTER),transform,series=rev(value2-c(0,cumsum(rev(value1[-1])))))
>
> but I get Table 2 instead of table 1.
>
> Thanks a lot for your attention and sorry for my english and my scary
> example....
>


-- 
Joris Meys
Statistical consultant

Ghent University
Faculty of Bioscience Engineering
Department of Applied mathematics, biometrics and process control

tel : +32 9 264 59 87
Joris.Meys at Ugent.be
-------------------------------
Disclaimer : http://helpdesk.ugent.be/e-maildisclaimer.php