jaredhuling / bigfastlm Goto Github PK

I have a data set of about a million rows.
variable c is a factor.

data <- model.matrix(~ poly(a,5,raw=TRUE) *  poly(b,5,raw=TRUE) + c)
bigMatrix <- data %>% bigmemory::as.big.matrix()
bigLmObj <- bigFastlm::bigLm(bigMatrix,y)
bigLmObj$coefficients %>% broom::tidy()

The output of this is as follows (all remaining names are ):

Have I done something wrong, or is there a bug where the covariate labels are dropped?

Thank you.

Hello,
I'm interested in the biglm method but I'm constantly getting this weird error message:
Error in bigLmPure(X, y, method, gigs, nslices) : std::bad_alloc Calls: system.time -> bigLm -> bigLm.default -> bigLmPure -> .Call

the code causing the error:
lm <- bigLm(X, y)
where X is a 'big.matrix' of bigMemory and y is of type 'numeric'.
Regards,
Kassim Bashir
[email protected]