injecting date column into time series ?

["Squid Splus" <squid.splus@gmail.com>]

I'm starting out with a small SAS data set and wanting to do some time series
analysis with Finmetrics but I'm not sure I have loaded and indexed the
data in the correctly for time series methods to work on it.


module(finmetrics);
mydata   <- importData("C:\\data\\mydata.sas7bdat", type="sas7");
ts    <- timeSeq(from="1/1/2002",to="3/1/2006", by = "months", + format = "%b %Y");


1. Warning message
sequence is over-specified -- ignoring length in: timeSeq(from =
    "1/1/2002", to = "3/1/2006", by = "months",  + format = "%b %Y" ....

mydata2 <- cbind(mydata,ts);

mytdatats <- timeSeries(mydata2);

mydatats     prints out   // date was a character variable in the SAS data set.


      
date         
x1        ..2 
 1 d2002Jan 0.016330678 01/01/2002
 2 d2002Feb 0.011731270 02/01/2002
 3 d2002Mar 0.008660340 03/01/2002
 
 
2. How do I address the column '..2' ?  As mydatats$..2  ?


3. I want to run time series methods from Zivots book, eg     
  
 kpss <- stationaryTest(mydatats$x1,trend="c");

Error Oject mydatats not found.

Do I need to attach to it if mydatats just found and printed it ?

Thanks for any suggestions.