[R] bigglm "update" with ff

[Thomas Lumley]

On Thu, 2 Apr 2009, Yue Sheng wrote:

> Hi, since bigglm doesn't have update, I was wondering how to achieve
> something like (similar to the example in ff package manual using biglm):

bigglm() doesn't have update because it needs to iterate over the data and 
update() couldn't do that.

Instead of calling update(), you have to supply a function that will load 
the next chunk of data, and that can be reset to start again at the 
beginning.  There is an example on the help page, and the code for 
bigglm.data.frame gives another example, as do the SQLiteConnection 
and RODBC methods.

 	-thomas



> first <- TRUE
> ffrowapply ({
>  if (first) {
>   first <- FALSE
>   fit <- bigglm(eqn, as.data.frame(bigdata[i1:i2,,drop=FALSE]), chunksize =
> 10000, family = binomial())
>  } else {
>   fit <- update(fit, as.data.frame(bigdata[i1:i2,,drop=FALSE]), chunksize =
> 10000, family = binomial())
>  }
> }, X=bigdata, VERBOSE = TRUE, BATCHSIZE = nmax)
>
> Many thanks.
> Yuesheng
>
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> 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.
>

Thomas Lumley			Assoc. Professor, Biostatistics
tlumley at u.washington.edu	University of Washington, Seattle