[R] sparse matrix and block of R

alessia matano alexis.rtd at gmail.com
Tue Jun 24 13:55:05 CEST 2008


Dear all,

I am writing you for two problems I can not solve with R. I used both
the 2.6.1 version and the 2.7.0 version.
I run a sintax written by a collegue of mine, where there are involved
both sparse matrix and quantile estimations. My data begins with
10008*14 observations and at maximum they reach 10008*834
observations.

My questions are the following:
first, when I give the command
X=as.matrix.csr(X)

and in matrix X there are some elements that repeat in some rows (i.e
in row 1 there is value 3, in row 2 there is again value 3 in the same
position), he gives me the following error:

Error in if (nnz == 0) { : valore mancante dove è richiesto TRUE/FALSE

(That means : missing value when it is required true/false).
When I take off that column with same value that repeats it works well
and creates the sparse matrix. Can you help me to correct this
problem? Is there a bug in the sparse matrix command?

My second problem is the following: after solving this problem I then
run a sintax where you run quantile regression repeatedly changing the
taus (the quantile estimated). Hence, when I run this program it stops
completely sending a microsoft message saying "An error has occurred,
the application will be closed" and there is no way to continue to run
the program. I thought the problem was the dimension of data, but
people told me they are not so huge. Can you help to understand this
problem? I tried also on other PC's and gives me the same problem. If
the problem is the memory, is there any way to increase it?
many thanks
alessia



More information about the R-help mailing list