R help - Feb 2002 - How to access objects outside an R function

Dear R-Users,

I have a dataframe (''forexdata'') of daily returns from the
foreign exchange market for three currencies - British Pound (bp), Canadian
Dollar(cd),
Deustche Mark (dm) vis-a-vis the US Dollar and the Date Of
Trade(yymmdd).

For some dates the returns are missing (recorded as zero) as there
were no trades in that currency for that date. My task is to
substitute the missing or zero values with the next non-zero value.

To this end I created two lag variables:

forexdata$counter <- rep(1:length(forexdata$DateOfTrade), 1);   

forexdata$bplag1 <- forexdata$bp[forexdata$counter+1];            
forexdata$bplag2 <- forexdata$bp[forexdata$counter+2];      

forexdata$counter <- NULL;
> forexdata;
   DateOfTrade DayOfWeek     bp     cd      dm bplag1 bplag2
1       730606         3 2.5715 1.0011 0.37700 2.5740 0.0000
2       730607         4 2.5740 1.0017 0.37890 0.0000 0.0000
3       730608         5 0.0000 1.0015 0.38170 0.0000 2.5747
4       730611         1 0.0000 1.0019 0.38000 2.5747 2.5740
5       730612         2 2.5747 1.0018 0.38290 2.5740 0.0000
6       730613         3 2.5740 1.0005 0.38025 0.0000 2.5805
7       730614         4 0.0000 0.9998 0.38560 2.5805 2.5815
8       730615         5 2.5805 1.0001 0.38840 2.5815 2.5725
9       730618         1 2.5815 1.0003 0.39100 2.5725     NA
10      730619         2 2.5725 1.0003 0.38960     NA     NA


I have written a function which uses two lag variables for
substitution. The substitution seems to be happening alright within
the function. However, as a last step, I am unable able to replace
these values in the original data frame
> exch.replace <- function (df, curr, currlag1, currlag2) {
                    if  ((df[curr]==0) && (df[currlag1]!=0))  ( 
                        df[curr] <- df[currlag1]
                    )
                    else if  ((df[curr]==0) && (df[currlag1]==0)) ( 
                        df[curr] <- df[currlag1] <- df[currlag2]
                    )                                  
                    return(df[curr] );                
                  }   
> apply(forexdata, 1, exch.replace,
                      curr="bp", 
                      currlag1="bplag1", 
                      currlag2="bplag2");       

The result of applying this function is:

      1      2      3      4      5      6      7      8      9     10 
 2.5715 2.5740 2.5747 2.5747 2.5747 2.5740 2.5805 2.5805 2.5815 2.5725
>
Any help in accomplishing this final step will be greatly appreciated.

a_rao		 


Find the best deals on the web at AltaVista Shopping!
http://www.shopping.altavista.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@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

anantha rao wrote:
> ...I have a dataframe ('forexdata') of daily returns from the
foreign exchange market for three currencies - British Pound (bp), Canadian
> Dollar(cd),
> Deustche Mark (dm) vis-a-vis the US Dollar and the Date Of
> Trade(yymmdd).
>
> For some dates the returns are missing (recorded as zero) as there
> were no trades in that currency for that date. My task is to
> substitute the missing or zero values with the next non-zero value.
Here is a function that will replace the zeros with the _last_  non-zero value
(which is what you did in your solution). Note that it will produce
NAs if the _first_ value in a column is zero. Supply the data frame and a vector
of column indices (in your case 3:5) to which you want to apply the
function. Hope this helps.

Jim


sub.last.value<-function(x.df,cols=1:dim(x.df)[2]) {
 nrows<-dim(x.df)[1]
 newcol<-vector("numeric",nrows)
 for(col in 1:length(cols)) {
  last.value<-NA
  for(row in 1:nrows) {
   if(x.df[row,cols[col]] > 0)
    newcol[row]<-last.value<-x.df[row,cols[col]]
   else newcol[row]<-last.value
  }
  x.df[cols[col]]<-newcol
 }
 return(x.df)
}


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._