Installing rgdal - backdate rgdal version versus update gdal version, and how?
What finally worked for me was trying to install an earlier version (v1.2-20) of rgdal
that was compatible with gdal 1.11.3
.
I was able to identify the correct version of rgdal using the documentation page here:
https://www.rdocumentation.org/packages/rgdal/versions/1.2-20
Then I used the devtools
package to install the correct version:
require(devtools)
install_version("rgdal", version="1.2-20")
This page has helpful information on installation of earlier versions of R packages:
https://support.rstudio.com/hc/en-us/articles/219949047-Installing-older-versions-of-packages
I attempted installing more recent versions of gdal both from source and with my package manager but continued to get various errors when I got to the step of trying to install rgdal
.
What I don't know is why I kept getting errors when trying to work with gdal upgrades.
Error: gdal-config not found while installing R dependent packages whereas gdal is installed
You need the -dev
package with headers and shared library links for development besides the normal package for deployment. Two different use cases as far as the distro is concerned.
On my distribution:
edd@max:/tmp$ dpkg -l | grep gdal | cut -c-72
ii libgdal1-1.7.0 1.7.3-6ubuntu3
ii libgdal1-dev 1.7.3-6ubuntu3
edd@max:/tmp$
and R CMD INSTALL rgdal_0.7-8.tar.gz
works just fine as you would expect from a CRAN package given all the build-time checks at CRAN.
Update late 2016: As @JoshO'Brien points out in the comment
Minor update: Here in 2016, running Ubuntu 14.04.2,
libgdal1h
seems to
have replacedlibgdal1
(thoughlibgdal1-dev
is still needed). At least
I get an error to that effect when I attempted toapt-get install libgdal1
For Ubuntu 16.04 the corresponding line would besudo apt-get install libgdal1i
Such renaming of upstream libraries is common; something such as apt-cache search libgdal
can help locate the current package names. The important key though is that the "abstract" development package libgdal-dev
is all that is needed to build as it pulls the "concrete" current run-time package (here: libgdal1i
) in via a dependency.
rgdal installed in Ubuntu 18.04 LTS, but R 3.5 not recognized it
The solution was to uninstall QGIS, and install GDAL first and re-install QGIS after.
Finally, the rGDAL was possible to be installed in R.
r Rgdal install fails on ubuntu 20.04 with error double red or corruption (!prev)
Following the directions at this link solved the rgdal problem. Now have a different, non-gdal error with sf so need to research that.
sf R package is not compatible with GDAL versions below 2.0.0 after installing gdal 2.3 from conda
Errors are caused by non-conda package installations. From a clean environment, the sf
package can be installed directly from conda
:
conda create -n rsf -c conda-forge r-sf
source activate rsf
Or from a clean environment:
conda install -c conda-forge r-sf
Related Topics
Download File from Internet via R Despite The Popup
How to Change Color Scheme in Corrplot
R Finding Duplicates in One Column and Collapsing in a Second Column
Ggplot and Axis Numbers and Labels
Making Commandargs Comma Delimited or Parsing Spaces
R: How to Filter a Timestamp by Hour and Minute
Split Line by Multiple Points Using Sf Package
Shiny Datatable in Landscape Orientation
What's a Prettier Way to Print Info with R
How to Programmatically Create Binary Columns Based on a Categorical Variable in Data.Table
Ifelse Assignment in Data.Table
R: Xmleventparse with Large, Varying-Node Xml Input and Conversion to Data Frame
Using Sample() with Sample Space Size = 1
How to Do Histograms of This Row-Column Table in R Ggplot