R help - Dec 2012 - Does R have an equivalent for Matlab's cell array?

Hi

I have a time series of measurements: every 10 min. a value was logged.

The data look like:

20100914 08:01     3.74
20100914 08:11     3.74
20100914 08:21     3.71
20100914 08:31     4.39
20100914 08:41     3.74

This data spans several months.

I would like to group the data per day. In Matlab it is fairly easy to 
obtain a cell array, of which the first column contains the date of each 
day, and the second column contains a vector, that in its turn contains 
all values for the corresponding day.

That looks like

14-9-2010    [3.74 3.74 3.71 4.39 3.74...]
15-9-2010    [...]
and so on.

Is it possible to create a similar data structure in in R? I was 
thinking of a data frame, similar to the Matlab's cell array, but it 
doesn't look like data frames can contain vectors. Something like a two 
dimensional list would do the trick, I believe, but does that exist?

Thanks

Kind regards

Matthijs Daelman

Data frames are lists of vectors. You can have the vectors in a data frame be of
mode list, if you wish. It would be more conventional to simply make the list
and not put it into the data frame, since columns of mode list give write.table
fits.

I prefer to simply keep the first form of your data and augment it with a Day
column. Then if I want information on a per-day basis I use aggregate from base
or ddply from the plyr package.

Although I prefer your first form (long format) there are many cases where it
makes sense to put groups of data values in separate columns (wide format). You
can use the reshape2 package for transforming back and forth.

You may also want to check out the Time Series task view on CRAN. Most of those
classes work with data in array representation for high speed.
---------------------------------------------------------------------------
Jeff Newmiller                        The     .....       .....  Go Live...
DCN:<jdnewmil at dcn.davis.ca.us>        Basics: ##.#.       ##.#.  Live
Go...
                                      Live:   OO#.. Dead: OO#..  Playing
Research Engineer (Solar/Batteries            O.O#.       #.O#.  with
/Software/Embedded Controllers)               .OO#.       .OO#.  rocks...1k
--------------------------------------------------------------------------- 
Sent from my phone. Please excuse my brevity.

Matthijs Daelman <matthijs.daelman at gmail.com> wrote:
>Hi
>
>I have a time series of measurements: every 10 min. a value was logged.
>
>The data look like:
>
>20100914 08:01     3.74
>20100914 08:11     3.74
>20100914 08:21     3.71
>20100914 08:31     4.39
>20100914 08:41     3.74
>
>This data spans several months.
>
>I would like to group the data per day. In Matlab it is fairly easy to 
>obtain a cell array, of which the first column contains the date of
>each 
>day, and the second column contains a vector, that in its turn contains
>
>all values for the corresponding day.
>
>That looks like
>
>14-9-2010    [3.74 3.74 3.71 4.39 3.74...]
>15-9-2010    [...]
>and so on.
>
>Is it possible to create a similar data structure in in R? I was 
>thinking of a data frame, similar to the Matlab's cell array, but it 
>doesn't look like data frames can contain vectors. Something like a two
>
>dimensional list would do the trick, I believe, but does that exist?
>
>Thanks
>
>Kind regards
>
>Matthijs Daelman
>
>______________________________________________
>R-help at r-project.org mailing list
>https://stat.ethz.ch/mailman/listinfo/r-help
>PLEASE do read the posting guide
>http://www.R-project.org/posting-guide.html
>and provide commented, minimal, self-contained, reproducible code.

On Dec 22, 2012, at 2:58 AM, Matthijs Daelman wrote:
> Hi
> 
> I have a time series of measurements: every 10 min. a value was logged.
> 
> The data look like:
> 
> 20100914 08:01     3.74
> 20100914 08:11     3.74
> 20100914 08:21     3.71
> 20100914 08:31     4.39
> 20100914 08:41     3.74
> 
> This data spans several months.
> 
> I would like to group the data per day. In Matlab it is fairly easy to
obtain a cell array, of which the first column contains the date of each day,
and the second column contains a vector, that in its turn contains all values
for the corresponding day.
> 
> That looks like
> 
> 14-9-2010    [3.74 3.74 3.71 4.39 3.74...]
> 15-9-2010    [...]
> and so on.
> 
> Is it possible to create a similar data structure in in R? I was thinking
of a data frame, similar to the Matlab's cell array, but it doesn't look
like data frames can contain vectors.
That's a surprising conclusion. Vectors are typical components of
dataframes.
> Something like a two dimensional list would do the trick, I believe, but
does that exist?
A dataframe _is_ a two-dimensional list.

You should probably explain what you are trying to do, since there is very
possibly a better way of structuring this problem without recasting the data
into a "wide format". Most of the data processing in R is handled very
well without such a strategy.

If you wanted to proceed despite my misgivings, then there are several functions
such as `split`, `tapply` and `by` that could allow you to segregate such data.
It's best to use dput to present the data since console output is often
ambiguous as to the underlying data structures.
From what you have offered it's not clear that you have yet done the data
input processing.

-- 

David Winsemius
Alameda, CA, USA

Hello,

Something like this?


dat <- read.table(text="
20100914 08:01     3.74
20100914 08:11     3.74
20100914 08:21     3.71
20100914 08:31     4.39
20100914 08:41     3.74
")
dat

split(dat$V3, dat$V1)


This gives you a list where each element is a (1-dimensional) vector, 
and each of those vectors can have any length, not necessarily the same.

Hope this helps,

Rui Barradas
Em 22-12-2012 10:58, Matthijs Daelman escreveu:
> Hi
>
> I have a time series of measurements: every 10 min. a value was logged.
>
> The data look like:
>
> 20100914 08:01     3.74
> 20100914 08:11     3.74
> 20100914 08:21     3.71
> 20100914 08:31     4.39
> 20100914 08:41     3.74
>
> This data spans several months.
>
> I would like to group the data per day. In Matlab it is fairly easy to 
> obtain a cell array, of which the first column contains the date of 
> each day, and the second column contains a vector, that in its turn 
> contains all values for the corresponding day.
>
> That looks like
>
> 14-9-2010    [3.74 3.74 3.71 4.39 3.74...]
> 15-9-2010    [...]
> and so on.
>
> Is it possible to create a similar data structure in in R? I was 
> thinking of a data frame, similar to the Matlab's cell array, but it 
> doesn't look like data frames can contain vectors. Something like a 
> two dimensional list would do the trick, I believe, but does that exist?
>
> Thanks
>
> Kind regards
>
> Matthijs Daelman
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide 
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

HI,

May be this helps:
dat1 <- read.table(text="
20100914 08:01,??? 3.74
20100914 08:11,??? 3.74
20100914 08:21,??? 3.71
20100914 08:31,??? 4.39
20100914 08:41,??? 3.74
20100915 08:01,??? 3.64
20100915 08:11,??? 3.54
20100915 08:21,??? 3.61
20100915 08:31,??? 4.49
20100915 08:41,??? 3.84
", sep=",",header=FALSE,stringsAsFactors=FALSE)
tapply(dat1$V2,list(as.Date(dat1$V1,format="%Y%m%d
%H:%M")),function(x) x)
#$`2010-09-14`
#[1] 3.74 3.74 3.71 4.39 3.74
#
#$`2010-09-15`
#[1] 3.64 3.54 3.61 4.49 3.84
A.K.




----- Original Message -----
From: Matthijs Daelman <matthijs.daelman at gmail.com>
To: r-help at r-project.org
Cc: 
Sent: Saturday, December 22, 2012 5:58 AM
Subject: [R] Does R have an equivalent for Matlab's cell array?

Hi

I have a time series of measurements: every 10 min. a value was logged.

The data look like:

20100914 08:01? ?  3.74
20100914 08:11? ?  3.74
20100914 08:21? ?  3.71
20100914 08:31? ?  4.39
20100914 08:41? ?  3.74

This data spans several months.

I would like to group the data per day. In Matlab it is fairly easy to obtain a
cell array, of which the first column contains the date of each day, and the
second column contains a vector, that in its turn contains all values for the
corresponding day.

That looks like

14-9-2010? ? [3.74 3.74 3.71 4.39 3.74...]
15-9-2010? ? [...]
and so on.

Is it possible to create a similar data structure in in R? I was thinking of a
data frame, similar to the Matlab's cell array, but it doesn't look like
data frames can contain vectors. Something like a two dimensional list would do
the trick, I believe, but does that exist?

Thanks

Kind regards

Matthijs Daelman

______________________________________________
R-help at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.