projectRaster 消耗太多内存 [英] projectRaster consuming too much memory
问题描述
我正在 R 中做一些空间方面的事情，突然之间我的一些代码将不再在我使用多年的计算机上运行，特别是因为它内存不足".
I'm doing some spatial stuff in R and out of the blue some of my code will no longer work on a computer that I've been using for years, specifically because it's "running out of memory."
## Raster going in
xx <- raster(fatNames[[i]])
xx
class : RasterLayer
dimensions : 5160, 14436, 74489760 (nrow, ncol, ncell)
resolution : 0.008333333, 0.008333333 (x, y)
extent : -172.3, -52, 23.5, 66.5 (xmin, xmax, ymin, ymax)
coord. ref. : +proj=longlat +datum=WGS84 +no_defs +ellps=WGS84 +towgs84=0,0,0
data source : D:\Dropbox\wintor_aux\Results\myluFatReq_100_2_fat.inf.tif
names : myluFatReq_100_2_fat.inf
values : 0.1831425, 3.449095 (min, max)
## transform
x.rast <- projectRaster(xx,crs = CRS("+init=epsg:2955"))
错误:无法提供足够的 4.2Gb 向量.
Error: can not adequate vector of 4.2Gb.
现在我知道这是一个大小合适的光栅(输入光栅大小:75.5 MB)但是机器有 32Gb 的 RAM，没有其他程序在运行，我仍然占用内存.多年来，我已经做了数百次这样的事情，据我所知，自从我大约一个月前做这件事以来，机器上的任何东西都没有改变.
Now I know this is a decent sized raster (Input raster size: 75.5 MB) however the machine has 32Gb of RAM, no other programs are running and I still swamp the memory. I've done things like this hundreds of times over the years and to the best of my knowledge nothing on the machine has changed since I did this last about a month ago.
我已经尝试重新启动会话、重新启动计算机并在谷歌上搜索了一下，但我仍然没有任何运气试图追踪问题.当我在具有 1/2 RAM 的机器上尝试相同的代码块时，我可以毫无问题地运行它(最大使用 ~ 8Gb ram).有人对我如何解决这个问题有任何建议吗?
I've tried restarting the session, restarting the computer and have googled around a bit but I still haven't had any luck in trying to trace down the issues. When I try the same code block on a machine with 1/2 the RAM I'm able to run it no problem (using ~ 8Gb ram max). Does anyone have any suggestions as to how I can chase this down?
> sessionInfo()
R version 3.5.1 (2018-07-02)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows >= 8 x64 (build 9200)
Matrix products: default
locale:
[1] LC_COLLATE=English_New Zealand.1252 LC_CTYPE=English_New Zealand.1252
[3] LC_MONETARY=English_New Zealand.1252 LC_NUMERIC=C
[5] LC_TIME=English_New Zealand.1252
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] raster_2.6-7 rgdal_1.3-3 sp_1.3-1 sf_0.6-3
[5] RevoUtils_11.0.1 RevoUtilsMath_11.0.0
loaded via a namespace (and not attached):
[1] Rcpp_0.12.18 lattice_0.20-35 class_7.3-14 grid_3.5.1 spData_0.2.9.0
[6] DBI_1.0.0 magrittr_1.5 e1071_1.7-0 units_0.6-0 tools_3.5.1
[11] yaml_2.2.0 compiler_3.5.1 classInt_0.2-3
推荐答案
您应该首先将栅格包更新为其当前 (CRAN) 版本.这可能会解决这个问题.如果不是，这可能是一种极端情况，您应该能够通过将内存参数设置为较低的值来解决该问题.在当前版本中，您可以执行 rasterOptions(memfrac=.3).在当前和旧版本中，您可以尝试 rasterOptions(maxmemory=1e+08)
You should first update the raster package to its current (CRAN) version. That might fix this. If not, this is probably a corner-case that you should be able to work-around by setting the memory parameters to a lower value. In the current version you could do rasterOptions(memfrac=.3). In the current and older versions you may try rasterOptions(maxmemory=1e+08)
顺便说一下，鉴于输入数据的经度跨度为 (172.3- 52) = 120 度 --- UTM 适用于大约 6 度的跨度.
By the way, your transformation to +proj=utm +zone=11 (please do not use espg codes in examples) seems rather questionable given that the input data has a longitude span of (172.3 - 52) = 120 degrees --- UTM is valid for a span of about 6 degrees.
这篇关于projectRaster 消耗太多内存的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
