Hi,
I am using version 2.1-1 of the ff package.
I have a data set with 80 million rows and I need to create a new ffdf
object, subseting by values in one of the original ffdf's columns. Here is
my code:
bigData <-
read.table.ffdf(file="/data/demodata/data/smallData.txt",
next.rows=1e5, head=TRUE, sep="|")
dim(bigData)
N <- nrow(bigData);N
select <- ff( vmode='logical', length=N, update=FALSE)
for (i in chunk(bigData, from=1, to=N, by=1e5))
select[i] <- ifelse(bigData[i,]$C.2 > 0.5, TRUE, FALSE)
bd <- bigData[select,]
Error in UseMethod("as.hi") :
no applicable method for 'as.hi' applied to an object of class
"c('ff_vector', 'ff')"
I tried something like the example in ?as.hi but I can not convert the
select vector into a RAM object and then use as.hi() b/c I do not have
enough RAM. What are my options for subseting this ffdf object by the
values of one of its columns?
More generally, is there a conversion guide / vignette showing how to
convert standard operations done on RAM data frames to operations done on
ffdf objects?
Here is my R version information> R.Version()
$platform
[1] "x86_64-pc-linux-gnu"
$`svn rev`
[1] "50720"
$version.string
[1] "R version 2.10.1 (2009-12-14)"
Thank you for your help.
OS
[[alternative HTML version deleted]]