Skip to content
Prev 43549 / 398506 Next

memory problem for R --Summary

Thank you very much for the replies you have sent me regarding the memory
problem.
The following is the summary
(I tried to read all the messages through. I apologized if I overlooked your
message)

Cheers,

Yun-Fang
----------------------------
Backgrounds:
a. Data: 1million rows with 73 numeric attributes
b. Environment: R 1.7.1 on FreeBSD 4.3 with  2GB memory and double CPU
   Pentium III/Pentium III Xeon/Celeron
    with  data seg size (kbytes) =1572864  limit

Suggested Solutions:
z. use SAS since SAS is not trying to read all the data into RAM.
a. random sampling from the large data set i.e. 10% of 1 million rows
    (the option singular.ok=TRUE can be used in lm for singular matrice.)
b. use kalman filter with migration variance =0. ( see the dse package for
details)
c. add the following configuration: options(object.size=1e8)
   Results:  still OOM
d. if data is all numeric, add colClasses="numeric" in read.table()
   Results: read.table read in the data successfully but I failed to access
the dataset after the loading
(even dataset[1:10,] didn't work)

----- Original Message -----
From: "Liaw, Andy" <andy_liaw at merck.com>
To: "'Yun-Fang Juan'" <yunfang at yahoo-inc.com>; "Prof Brian Ripley"
<ripley at stats.ox.ac.uk>
Cc: <r-help at stat.math.ethz.ch>
Sent: Friday, January 30, 2004 11:44 AM
Subject: RE: [R] memory problem for R
----
e-mail
----