On 18-May-13, at 8:14 AM, Robert J Goedman wrote:
Don,
Slightly modified, your example:
N <- 5e4
cat('Create myfile\n')
system.time(myfile <- data.frame(x1=runif(N),x2=runif(N),x3=runif(N)))
head(myfile)
cat('\nSave myfile\n')
system.time(save("myfile",file="./R/R_projects/R Support/Debug/
myfile.RData"))
remove(myfile)
cat('Load myfile\n')
system.time(load("./R/R_projects/R Support/Debug/myfile.RData"))
head(myfile)
cat('\nDump myfile\n')
system.time(dump("myfile",file="./R/R_projects/R Support/Debug/
myfile.R"))
remove(myfile)
cat('Source myfile\n')
system.time(source("./R/R_projects/R Support/Debug/myfile.R"))
head(myfile)
produces:
source('/Users/rob/Projects/R/R_projects/R support/Debug/
DumpAndSave_test.r', print.eval=TRUE)
Create myfile
user system elapsed
0.005 0.001 0.006
x1 x2 x3
1 0.611084090 0.4841179 0.47273937
2 0.596409008 0.2234015 0.19283019
3 0.273999461 0.7524927 0.02346372
4 0.691567956 0.4921780 0.48312317
5 0.001781206 0.1211497 0.12041949
6 0.002494183 0.6935539 0.23262610
Save myfile
user system elapsed
0.161 0.001 0.162
Load myfile
user system elapsed
0.007 0.001 0.008
x1 x2 x3
1 0.611084090 0.4841179 0.47273937
2 0.596409008 0.2234015 0.19283019
3 0.273999461 0.7524927 0.02346372
4 0.691567956 0.4921780 0.48312317
5 0.001781206 0.1211497 0.12041949
6 0.002494183 0.6935539 0.23262610
Dump myfile
user system elapsed
0.335 0.005 0.338
Source myfile
user system elapsed
22.252 0.238 22.440
x1 x2 x3
1 0.611084090 0.4841179 0.47273937
2 0.596409008 0.2234015 0.19283019
3 0.273999461 0.7524927 0.02346372
4 0.691567956 0.4921780 0.48312317
5 0.001781206 0.1211497 0.12041949
6 0.002494183 0.6935539 0.23262610
For N=1e5, the elapsed time for source() is ~100seconds. It is
slow, but eventually will return on my machine .
Regards,
Rob J. Goedman
[email protected]
Rob -- thanks for taking the time to try that out. Simon's
suggestion to switch to save() and load() seems better all around, so
I'll move to that usage. I should get out more -- that seems to
have been the standard for a while.
Don
On May 17, 2013, at 7:16 PM, Don McKenzie <[email protected]>
wrote:
Yes. Duh. Overthinking the "reproducible" part and responding too
quickly. Here's one that produces the behavior.
myfile <- data.frame(x1=runif(100000),x2=runif(100000),x3=runif
(100000))
dump("myfile",file="myfileR")
source("myfileR")
The source() part takes ~10 sec. in R 2.9.2 on my old powermac,
but hangs in 3.0.0.
Don
On 17-May-13, at 6:51 PM, Simon Urbanek wrote:
On May 17, 2013, at 9:41 PM, Donald McKenzie wrote:
Simon -- I can't reproduce it for the list because the window
it's running in always hangs and R freezes. I tried. But it
basically looks like
dump("my.file",file="myfileR")
source("myfileR")
HANG..... (yellow colored ball)
Well, you can simply provide the myfileR that hangs for you ...
Cheers,
Simon
I will go with your suggestions below and report back if need
be. Thanks for the help.
Don
On Fri, 17 May 2013, Simon Urbanek wrote:
On May 17, 2013, at 7:20 PM, Don McKenzie wrote:
I checked the archives since the release of R 3.0.0 and didn't
see anything on this.
Trying to import data previously exported with dump(), source
() hangs and I have to force quit R. This happens with data
objects
as small as 5 kb.
Do you have a reproducible example?
Has anyone encountered this and found a workaround? Is there
now a preferred method to export R objects?
The preferred method has always been to use serialization (save/
load or saveRDS/readRDS).
Cheers,
Simon
Thanks for any suggestions.
on a macbook pro with 4 gb RAM.
R.version
_
platform x86_64-apple-darwin10.8.0
arch x86_64
os darwin10.8.0
system x86_64, darwin10.8.0
status
major 3
minor 0.0
year 2013
month 04
day 03
svn rev 62481
language R
version.string R version 3.0.0 (2013-04-03)
nickname Masked Marvel
Don McKenzie
Research Ecologist
Pacific Wildland Fire Sciences Lab
US Forest Service
Affiliate Professor
School of Environmental and Forest Sciences
University of Washington
phone: 206-732-7824
[email protected]
_______________________________________________
R-SIG-Mac mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-mac
_______________________________________________
R-SIG-Mac mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-mac
Don McKenzie, Research Ecologist
Pacific Wildland Fire Sciences Lab
US Forest Service
phone: 206-732-7824
Affiliate Professor
School of Environmental and Forest Sciences
University of Washington
_______________________________________________
R-SIG-Mac mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-mac
Don McKenzie, Research Ecologist
Pacific Wildland Fire Sciences Lab
US Forest Service
phone: 206-732-7824
Affiliate Professor
School of Environmental and Forest Sciences
University of Washington
_______________________________________________
R-SIG-Mac mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-mac
