R help - Apr 2008 - Summarize data for MCA (FactoMineR)

Hi :-)

I'm new to R and I started using it for a project (I'm the CS guy in a
group
of statisticians helping them find out how to solve issues as they come out).
This is my first post to the list and I am starting to learn R.

Well, they were used to doing MCA analysis in other programs where the data
seems to be preprocessed automatically before running MCA.

So, they need to process a data set that comes with N=1000000 of elements,
but there are really about N/100 distinct elements over all the variables, so
the MCA can be run in reasonable time summarizing data.

So, the question is:

How can I turn x from:

x <-
structure(list(weight = c(1, 1, 2, 1, 2), var1 = structure(c(1L,
1L, 1L, 1L, 2L), .Label = c("A", "C"), class =
"factor"), var2 structure(c(1L,
1L, 1L, 1L, 2L), .Label = c("B", "D"), class =
"factor")), .Names = c("weight",
"var1", "var2"), row.names = c(NA, 5L), class =
"data.frame")

to:

y <-
structure(list(weihgt = c(5L, 2L), var1 = structure(1:2, .Label =
c("A",
"C"), class = "factor"), var2 = structure(1:2, .Label =
c("B",
"D"), class = "factor")), .Names = c("weihgt",
"var1", "var2"
), class = "data.frame", row.names = c(NA, -2L))

using R?

That is, from:
> x
  weight var1 var2
1      1    A    B
2      1    A    B
3      2    A    B
4      1    A    B
5      2    C    D

to:
> y
  weihgt var1 var2
1      5    A    B
2      2    C    D


The idea is that there is one occurrence of "A B" repeated 4 times in
the original table,
and it is summarized in the second table, computing the sum of the weights.

I solved the problem using Perl, but I'd like to know what I have to
read in order to
do it in R.

Regards,
Nelson.-

-- 
http://arhuaco.org

"Nelson Castillo" <nelsoneci at gmail.com> wrote in
news:2accc2ff0804251655o32686b99j73cf7df37243d08f at mail.gmail.com: 
> Hi :-)
> 
> I'm new to R and I started using it for a project (I'm the CS guy
in
> a group of statisticians helping them find out how to solve issues
> as they come out). This is my first post to the list and I am
> starting to learn R. 
> 
> Well, they were used to doing MCA analysis in other programs where
> the data seems to be preprocessed automatically before running MCA.
> 
> So, they need to process a data set that comes with N=1000000 of
> elements, but there are really about N/100 distinct elements over
> all the variables, so the MCA can be run in reasonable time
> summarizing data. 
> 
> So, the question is:
> 
> How can I turn x from:
> 
> x <-
> structure(list(weight = c(1, 1, 2, 1, 2), var1 = structure(c(1L,
> 1L, 1L, 1L, 2L), .Label = c("A", "C"), class =
"factor"), var2 > structure(c(1L,
> 1L, 1L, 1L, 2L), .Label = c("B", "D"), class =
"factor")), .Names > c("weight", "var1",
"var2"), row.names = c(NA, 5L), class > "data.frame")
> 
> to:
> 
> y <-
> structure(list(weihgt = c(5L, 2L), var1 = structure(1:2, .Label >
c("A", "C"), class = "factor"), var2 =
structure(1:2, .Label > c("B", "D"), class =
"factor")), .Names = c("weihgt", "var1",
"var2"
> ), class = "data.frame", row.names = c(NA, -2L))
> 
> using R?
> 
> That is, from:
> 
>> x
>   weight var1 var2
> 1      1    A    B
> 2      1    A    B
> 3      2    A    B
> 4      1    A    B
> 5      2    C    D
> 
> to:
> 
>> y
>   weihgt var1 var2
> 1      5    A    B
> 2      2    C    D
> 
Does this suffice?

s.wt <- with(x, 
          aggregate(weight, by=list(var1=var1,var2=var2), sum)
             )
#> s.wt
#  var1 var2 x
#1    A    B 5
#2    C    D 2

#then fix names
names(s.wt)[3] <- "weight"

#> s.wt
#  var1 var2 weight
#1    A    B      5
#2    C    D      2

I believe that the reshape or reShape packages could do this in one 
step.


-- 
David Winsemius

> 
> The idea is that there is one occurrence of "A B" repeated 4
times
> in the original table,
> and it is summarized in the second table, computing the sum of the
> weights. 
> 
> I solved the problem using Perl, but I'd like to know what I have to
> read in order to
> do it in R.
> 
> Regards,
> Nelson.-
>