[R-sig-Geo] Problem with writeOGR and overwrite_layer=T

Mon May 14 20:05:41 CEST 2012

On Mon, 14 May 2012, Roger Bivand wrote:

> On Mon, 14 May 2012, Bastien.Ferland-Raymond at mrnf.gouv.qc.ca wrote:
>
>> Dear list,
>> 
>> I have to produce shapefile for a coworker which uses ArcGIS and I've hit a 
>> problem with the writeOGR function from the rgdal package.
>
> I can reproduce the problem on Windows 7 with GDAL 1.9.0, that is, that parts 
> of shapefiles created by OGR (the *.dbf and *.shp files) cannot be deleted by 
> poDS->DeleteLayer() - line 473 in rgdal/src/ogrsource.cpp. This also applies 
> before trying to overwrite them - they cannot be deleted in Windows Explorer 
> either until the R session exits. Running as administrator does not help 
> either.
>
> This appears to be an interaction between Windows (7 in my case), GDAL >= 
> 1.8.0, which introduced bans on overwriting for Shapefiles, and possibly some 
> unknown factors.
>
> This is not a problem I can resolve, certainly not any time soon. It is not 
> an R code problem as such (I believe), and the stackoverflow speculation is 
> misleading, this has nothing to do with connections, only with Windows 
> apparent file usage status.
>
> I would appeal to interested Windows users to contribute a solution - you 
> will need to be able to cross-compile GDAL for Windows under MSYS, and to 
> build R packages.

Actually, it isn't so hard to contribute, as WinBuilder should be able to 
produce viable Windows binary packages. So contributors should check out 
rgdal anonymously from R-Forge, then let me have patches to insert, submit 
to WinBuilder, and return a download link (say 6 hours later) to see if 
the patch helped. So maybe a local build of GDAL may not be needed, but 
the time involved would be greater.

Roger

> If Windows users know how to unlink files with an incorrect 
> usage status without exiting R, this would possibly help, but doesn't get us 
> much further.
>
> Other GDAL/OGR uses such as ogr2ogr exit on error, but it might be helpful to 
> know if ogr2ogr does the same - that is, is unable to overwrite a layer. 
> MapInfo File, GML and other drivers do not seem to suffer from the same 
> issue. Again, checking across a wider range of drivers would help.
>
> Roger
>
>> 
>> Here is an example:
>> 
>> ###
>> 
>>> library(rgdal)
>> Loading required package: sp
>> Geospatial Data Abstraction Library extensions to R successfully loaded
>> Loaded GDAL runtime: GDAL 1.9.0, released 2011/12/29
>> Path to GDAL shared files: 
>> \\sebpriv/usagers/Ferba1/R/R-2.15.0/library/rgdal/gdal
>> Loaded PROJ.4 runtime: Rel. 4.7.1, 23 September 2009, [PJ_VERSION: 470]
>> Path to PROJ.4 shared files: 
>> \\sebpriv/usagers/Ferba1/R/R-2.15.0/library/rgdal/proj
>>> sessionInfo()
>> R version 2.15.0 (2012-03-30)
>> Platform: i386-pc-mingw32/i386 (32-bit)
>> 
>> locale:
>> [1] LC_COLLATE=French_Canada.1252  LC_CTYPE=French_Canada.1252 
>> LC_MONETARY=French_Canada.1252 LC_NUMERIC=C 
>> LC_TIME=French_Canada.1252
>> 
>> attached base packages:
>> [1] stats     graphics  grDevices utils     datasets  methods   base
>> 
>> other attached packages:
>> [1] rgdal_0.7-8 sp_0.9-98
>> 
>> loaded via a namespace (and not attached):
>> [1] grid_2.15.0    lattice_0.20-6
>>> setwd("C:\\Documents and Settings\\ferba1\\Bureau\\test rgdal")
>> 
>> # make a basic shape:
>> 
>>> coo <- 
>>> coordinates(cbind(runif(100,220000,300000),runif(100,5160000,5220000)))
>>> point.shape <- 
>>> SpatialPointsDataFrame(coo,data.frame(ID=paste("point",1:100,sep="_")),
>> +   proj4string = CRS(" +proj=tmerc +lat_0=0 +lon_0=-73.5 +k=0.9999 
>> +x_0=304800 +y_0=0 +datum=NAD83 +units=m +no_defs +ellps=GRS80 
>> +towgs84=0,0,0"))
>> 
>> # now if I try to write itm it works fine for the first 2 shapes:
>> 
>>> writeOGR(point.shape,".","point_shape1", driver="ESRI Shapefile")
>>> 
>>> writeOGR(point.shape,".","point_shape2", driver="ESRI Shapefile" )
>> 
>> # but then if I try to overwrite, it fails:
>> 
>>> writeOGR(point.shape,".","point_shape1", driver="ESRI Shapefile", 
>>> overwrite_layer=T)
>> Error in writeOGR(point.shape, ".", "point_shape1", driver = "ESRI 
>> Shapefile",  :
>>  Creation of output file failed
>> 
>> # Interestingly, it erases 2 files from point_shape1 before failing (the 
>> .prj and the .shx).
>> # and from now on, it's impossible to create new shapefile:
>> 
>>> writeOGR(point.shape,".","point_shape3", driver="ESRI Shapefile" )
>> Error in writeOGR(point.shape, ".", "point_shape3", driver = "ESRI 
>> Shapefile") :
>>  Creation of output file failed
>> ###
>> 
>> However, after restarting the whole session and erasing the files created, 
>> and try this new order:
>> 
>> ###
>> 
>>> 
>>> writeOGR(point.shape,".","point_shape1", driver="ESRI Shapefile")
>>> 
>>> writeOGR(point.shape,".","point_shape1", driver="ESRI Shapefile", 
>>> overwrite_layer=T)
>>> 
>>> writeOGR(point.shape,".","point_shape2", driver="ESRI Shapefile" )
>>> 
>>> writeOGR(point.shape,".","point_shape3", driver="ESRI Shapefile" )
>>> 
>> 
>> ###
>> Everything works fine.  So two questions arise:
>> 
>> 1) why does overwrite_layer doesn't work when the order is changed, can it 
>> be fixed?
>> 2) How to "reset" the writeOGR function after a fail as it seems once it 
>> fails, it get impossible to reuse it in the same session.
>> 
>> After reading this post: 
>> http://stackoverflow.com/questions/6219368/problem-writing-shapefiles-with-writeogr-in-r, 
>> I think it may be some connection issue.  Is there a easy way to close 
>> manually the connection after using writeOGR?  Something like 
>> close.OGR.connection()?
>> 
>> I know maptools::writeSpatialShape can also be used to write shapefile, 
>> however it doesn't write the projection information which I want. It seems 
>> to me that there is a problem with writeOGR, which is a good function so it 
>> could deserve to be fix.
>> 
>> Thanks for your help,
>> 
>> Bastien Ferland-Raymond, M.Sc. Stat., M.Sc. Biol.
>> Division des orientations et projets spéciaux
>> Direction des inventaires forestiers
>> Ministère des Ressources naturelles et de la Faune
>> 
>> _______________________________________________
>> R-sig-Geo mailing list
>> R-sig-Geo at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>> 
>
>

-- 
Roger Bivand
Department of Economics, NHH Norwegian School of Economics,
Helleveien 30, N-5045 Bergen, Norway.
voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no