R help - Jul 2006 - which data structure for a set of time series ?

Hello,
I'm a R newcomer and I'm wondering the kind of data structure that would
best fit to my problem:
my data are equities (stocks) : so I have a time serie (say 1 year of weekly
data), and a bunch of qualitative + quantitative variables :
the sector of the stock (biotech/finance...), the geographical region, the
name, ISIN code, P/E ratios, whatever...

The data.frame is perfect for the qualitative variables : a row is a stock,
a column is a variable .
But where can I store the time serie? If I put each weekly data in a column
of the data frame
, I'm not sure that it's the best structure when I will want to test
some models on these
(regressions/arma/whatever).
I would prefer to have a new data frame, where each column is a stock, each
row a date, to do the regressions (across time more easily)

So I would end up with 2 data frames, one of each transposed... No good,
probably.

I thought of a set of 2 objects : 1 data frame (for the qualitative
variables), and one list of time series, in the same order, with the risk of

forgetting to subset one objet and not the other, etc...

Do you happen to have the same kind of problem?
Is there something equivalent in R as the matlab 'structures' ? (kinda R
data frame, where you can store multi-dimensional objects in a column)

Many thanks in advance.

Regards

Roolio

	[[alternative HTML version deleted]]

A data.frame is perfect for storing both qualitative and quantitative 
variables together when they all have more or less the same length and 
where the observations are connected by something like time.  Without 
substantive common connection, people usually put variables in different 
data.frames.

	  Have you reviewed the information on data.frames in "An Introduction 
to R", available via 'help.start()'?  Also regarding coding time,
have
you reviewed the "zoo" vignette?  (See, e.g., 
"http://finzi.psych.upenn.edu/R/Rhelp02a/archive/67006.html".)

	  Hope this helps.
	  Spencer Graves

Roolio wrote:
> Hello,
> I'm a R newcomer and I'm wondering the kind of data structure that
would
> best fit to my problem:
> my data are equities (stocks) : so I have a time serie (say 1 year of
weekly
> data), and a bunch of qualitative + quantitative variables :
> the sector of the stock (biotech/finance...), the geographical region, the
> name, ISIN code, P/E ratios, whatever...
> 
> The data.frame is perfect for the qualitative variables : a row is a stock,
> a column is a variable .
> But where can I store the time serie? If I put each weekly data in a column
> of the data frame
> , I'm not sure that it's the best structure when I will want to
test
> some models on these
> (regressions/arma/whatever).
> I would prefer to have a new data frame, where each column is a stock, each
> row a date, to do the regressions (across time more easily)
> 
> So I would end up with 2 data frames, one of each transposed... No good,
> probably.
> 
> I thought of a set of 2 objects : 1 data frame (for the qualitative
> variables), and one list of time series, in the same order, with the risk
of
> 
> forgetting to subset one objet and not the other, etc...
> 
> Do you happen to have the same kind of problem?
> Is there something equivalent in R as the matlab 'structures' ?
(kinda R
> data frame, where you can store multi-dimensional objects in a column)
> 
> Many thanks in advance.
> 
> Regards
> 
> Roolio
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html