Hello,
Is there a command for freeing up the memory used by R in holding data
tables?
The structure of the procedure I have is as follows:
1) Read multiple txt files in using read.table(...).
2) Combine the read tables using rbind(...).
3) Attach the data using attach(...) and then use do a multiple regression
using lm(...).
So far so good, but when I then perform a further regression by taking out
factors, I run into memory issues, getting warnings such as:
"1: In as.list.data.frame(X) :
Reached total allocation of 1535Mb: see help(memory.size)"
As it stands, I have to close and then restart R, read in the same data
again and run with the new reduced number of factors.
My thinking was that, if I could reclaim the memory held by the already read
data files, keeping only the result of the rbind process, I could avoid the
duplication. I have therefore tried (very amateurishly) to reset the read
data to zero using:
Read_data_1=(0)
Read_data_2=(0)... etc
Followed by:
gc()
However this doesn't get solve the problem. Is there a better way of
getting R to "forget" the data tables it was holding and free up the memory?
For info: I am also specifying colClasses when first reading the data in, to
try to make it more memory-efficient (following:
http://www.biostat.jhsph.edu/~rpeng/docs/R-large-tables.html).
Other alternatives are trying the 3GB switch (XP Home, with 4GB RAM).
Another alternative is trying to use the sqldf package to bring the data in,
which one poster very helpfully suggested in response to an earlier
question. I may end up trying that, but as I have not used SQL, I am a
little daunted by the prospect.
I would really appreciate any suggestions. Thanks.
Guy Green
--
View this message in context:
http://www.nabble.com/Clearing-out-or-reclaiming-memory-tp24268680p24268680.html
Sent from the R help mailing list archive at Nabble.com.
______________________________________________
R-help@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.
