[[Image(source:EDA/data/Docs/trac/gdal.jpg)]] = Using GDAL - Geospatial Data Abstraction Library and rgdal = ---- back to Recipes for EDA ["CookBook Eda"] = gdal = ---- gdal url:[http://www.gdal.org/][[BR]] is a translator library for raster geospatial data formats that is released under an X/MIT style Open Source license by the Open Source Geospatial Foundation. As a library, it presents a single abstract data model to the calling application for all supported formats. It also comes with a variety of useful commandline utilities for data translation and processing. [[BR]] '' Cédric => What I've found at this site is not really helpfull, it is only python and c++ code. ''[[BR]] url:[http://www.bostongis.com/?content_name=ogr_cheatsheet] => '' Cedric: very good site ! '' [[BR]] url:[http://trac.osgeo.org/gdal/] '' Cedric : not really usefull '' ''' Below we suppose that you have the Qgis tool installed, so your path is C:\OSGeo4W ''' Qgis has gdal included. Gdal comes with two main tools (executable files), the files for gdal are in {{{ C:\OSGeo4W\share\gdal }}} call >cmd {{{ C:\OSGeo4W\bin\ogr2ogr }}} or {{{ C:\OSGeo4W\bin\ogrinfo }}} === Translating .mdb into shapes outside ESRI === ---- {{{ C:\OSGeo4W\bin\ogr2ogr -f "ESRI Shapefile" C:\base\basesig\Czhyd.shp C:\base\basesig\Czhyd.mdb }}} '' Cedric : this one works but warning there should be no space in the path (ex : no "document and settings/"...) '' === Conversion from PostGIS to ESRI Shape === '' adapted from url: ''[http://www.bostongis.com/?content_name=ogr_cheatsheet] The '''pgsql2shp''' and '''shp2pgsql''' are usually the best tools for converting back and forth between PostGIS and ESRI for 2 main reasons. * ''It has fewer idiosyncracies when converting data.'' * ''It has a lot fewer dependencies so can fit on your floppy.'' If you really want to use Ogr2Ogr for this kind of conversion, below is the standard way to do it Here the example should export the table riversegments[[BR]] Cedric {{{ C:\OSGeo4W\bin\ogr2ogr -f "ESRI Shapefile" C:\base\basesig\essai.shp PG:"host=localhost user=postgres dbname=CCM password=postgres port=5433" "riversegments" }}} Celine (pwd to be precised) {{{ C:\OSGeo4W\bin\ogr2ogr -f "ESRI Shapefile" C:\base\basesig\essai.shp PG:"host=localhost user=postgres dbname=CCM password=PWD" "riversegments" }}} '' Cedric : celui ci marche mais j'ai eu du mal, ne marche pas une deuxième fois si les shape sont déjà crées dans le répertoire ''[[BR]] {{{ C:\OSGeo4W\bin\ogr2ogr -f "ESRI Shapefile" C:\base\basesig\essai PG:"host=localhost user=postgres dbname=CCM password=postgres port=5433" "riversegments" }}} '' Cedric : OK marche de nouveau ''[[BR]] '''Selecting specific fields, sets of data and Geometry'''[[BR]] Sometimes you have more than one geometry field in a table, and ESRI shape can only support one geometry field per shape. Also you may only want a subset of data. In these cases, you will need to select the geometry field to use. The most flexible way to do this is to use the -sql command which will take any sql statement. {{{ C:\OSGeo4W\bin\ogr2ogr -f "ESRI Shapefile" C:\base\basesig\essai1 PG:"host=locahost user=postgres dbname=CCM password=postgres port=5433" -sql "SELECT gid, the_geom FROM riversegments" }}} '' Cedric : ne marche pas, j'obtiens un message du type Unable to open datasource with following drivers, celui là on l'a très très souvent, j'ai essayé plusieurs trucs je ne comprends pas, la requete marche dans pgadmin''[[BR]] ''' One way in which ogr2ogr excels above using the pgsql2shp tool is that ogr2ogr can export multiple tables at once. This is pretty handy for sharing your postgis data with others who do not have a postgis database. ''' The code below will export all your postgis tables out into a folder called mydatadump in ESRI shape (shp) format. {{{ C:\OSGeo4W\bin\ogr2ogr -f "ESRI Shapefile" essai2 PG:"host=locahost user=postgres dbname=CCM password=postgres port=5433" }}} Now most of the time you probably only want to output a subset of your postgis tables rather than all your tables. This code exports only the riversegments and rivernodes tables to a folder called c:/base/basesig/mydatadump in ESRI shapefile format {{{ C:\OSGeo4W\bin\ogr2ogr -f "ESRI Shapefile" c:/base/basesig/mydatadump PG:"host=localhost user=postgres dbname=CCM password=postgres port=5433" riversegments rivernodes }}} '' Cedric Celui là marche même si il n'y a pas de guillemet" = Connecting PostGIS with R using Rgdal = -------------------------------------------------------------------- The rddal is made of functions which point out to the c code functions of gdal, it seems to have been developped for intamap eu project which is also an R library and the most interesting page can be found at the following location url: [http://wiki.intamap.org/index.php/PostGIS] et bien sûr ça marche pas ...'' [[BR]] A nice example script follows which I translated into the following code {{{ library(rgdal) meuse = readOGR("PG:dbname=CCM", "riversegments") }}} but it doesn't work So after a quite long time spent in gdal, I can translate into the more complex dsn {{{ require(rgdal) readOGR(dsn="PG:host=locahost user=postgres dbname=CCM password=postgres port=5433",layer="riversegments") }}} However the R invariably returns an absolutely not helpful code, and the same with R2.10 or R2.10.1 {{{ > require(rgdal) Le chargement a nécessité le package : rgdal Le chargement a nécessité le package : sp Geospatial Data Abstraction Library extensions to R successfully loaded Loaded GDAL runtime: GDAL 1.6.2, released 2009/07/31 Path to GDAL shared files: C:/Program Files/R/R-2.10.0/library/rgdal/gdal Loaded PROJ.4 runtime: Rel. 4.6.1, 21 August 2008 Path to PROJ.4 shared files: C:/Program Files/R/R-2.10.0/library/rgdal/proj Message d'avis : le package 'sp' a été compilé avec la version R 2.10.1 > readOGR(dsn="PG:host=localhost dbname=CCM user=postgres passwd=postgres port=5433",layer="riversegments") Erreur dans ogrInfo(dsn = dsn, layer = layer, input_field_name_encoding = input_field_name_encoding) : Cannot open file }}} Interestingly it is the pointer to ogrinfo() which fails so we will try this with ogrinfo (which is found at the same place than ogr2ogr {{{ C:\OSGeo4W\bin\ogrinfo -f "PostgreSQL" PG:"host=locahost user=postgres dbname=CCM password=postgres port=5433" }}} which returns {{{ C:\base\basesig>C:\OSGeo4W\bin\ogrinfo -f "PostgreSQL" PG:"host=locahost user=po stgres dbname=CCM password=postgres port=5433" Usage: ogrinfo [--help-general] [-ro] [-q] [-where restricted_where] [-spat xmin ymin xmax ymax] [-fid fid] [-sql statement] [-al] [-so] [--formats] datasource_name [layer [layer ...]] }}} There does not seem to be any problem and also nothing is returned ... The next example for writing {{{ library(rgdal) writeOGR(meuse[1:10,], "PG:dbname=postgis", "meuse3", "PostgreSQL") }}} returns {{{ Erreur dans writeOGR(meuse[1:10, ], "PG:dbname=postgis", "meuse3", "PostgreSQL") : No such driver: PostgreSQL }}} All examples which do not imply postgres work {{{ logo <- system.file("pictures/logo.jpg", package="rgdal")[1] x <- new("GDALReadOnlyDataset", logo) x getDriverLongName(getDriver(x)) displayDataset(x) ogrDrivers() dsn <- system.file("vectors", package = "rgdal")[1] ogrInfo(dsn=dsn, layer="cities") cities <- readOGR(dsn=dsn, layer="cities") summary(cities) ogrInfo(dsn=dsn, layer="kiritimati_primary_roads") kiritimati_primary_roads <- readOGR(dsn=dsn, layer="kiritimati_primary_roads") summary(kiritimati_primary_roads) ogrInfo(dsn=dsn, layer="scot_BNG") scot_BNG <- readOGR(dsn=dsn, layer="scot_BNG") summary(scot_BNG) if ("GML" %in% ogrDrivers()$name) { dsn <- system.file("vectors/airports.gml", package = "rgdal")[1] airports <- try(readOGR(dsn=dsn, layer="airports")) if (class(airports) != "try-error") summary(airports) } dsn <- system.file("vectors/ps_cant_31.MIF", package = "rgdal")[1] ogrInfo(dsn=dsn, layer="ps_cant_31") ps_cant_31 <- readOGR(dsn=dsn, layer="ps_cant_31") summary(ps_cant_31) dsn <- system.file("vectors/Up.tab", package = "rgdal")[1] ogrInfo(dsn=dsn, layer="Up") Up <- readOGR(dsn=dsn, layer="Up") summary(Up) dsn <- system.file("vectors/test_trk2.gpx", package = "rgdal")[1] test_trk2 <- try(readOGR(dsn=dsn, layer="tracks")) if (class(test_trk2) != "try-error") summary(test_trk2) test_trk2pts <- try(readOGR(dsn=dsn, layer="track_points")) if (class(test_trk2pts) != "try-error") summary(test_trk2pts) dsn <- system.file("vectors", package = "rgdal")[1] ogrInfo(dsn=dsn, layer="trin_inca_pl03") birds <- readOGR(dsn=dsn, layer="trin_inca_pl03") summary(birds) }}} mmmh ... and PostgreSQL in not listed in ogrDrivers() {{{ ogrDrivers() }}} {{{ > ogrDrivers() name write 1 AVCBin FALSE 2 AVCE00 FALSE 3 BNA TRUE 4 CSV TRUE 5 DGN TRUE 6 ESRI Shapefile TRUE 7 Geoconcept TRUE 8 GeoJSON TRUE 9 GML TRUE 10 GMT TRUE 11 GPX TRUE 12 KML TRUE 13 MapInfo File TRUE 14 Memory TRUE 15 REC FALSE 16 S57 TRUE 17 SDTS FALSE 18 TIGER TRUE 19 UK .NTF FALSE 20 VRT FALSE 21 XPlane FALSE }}} in url:[http://cran.r-project.org/web/packages/rgdal/index.html] rgdal.pdf read.ogr details we read :[[BR]] "The drivers available will depend on the installation of GDAL/OGR, and can vary; the ogrDrivers() function shows which are available, and which may be written (but all are assumed to be readable). Note that stray files in data source directories (such as *.dbf) may lead to suprious errors that ac- companying *.shp are missing. [[BR]]" --------------------------- After sending a mail to the R-Sig-Geo list, the answer is as follows (by Roger Bivand) If you mean the Windows binary rgdal package, then the answer is in {{{ file.show(system.file("README", package="rgdal")) file.show(system.file("README.windows", package="rgdal")) }}} The Windows binary package builds the minimal set of drivers, with only one external dependency (expat for reading GPX and KML). The file refered to describes how you might build rgdal from source, using FWTools or OSGEO4W as the provider of GDAL and its dependent DLLs. If you do try this out (look for the chunk beginning: "Initial notes for OSGeo4W"), and if: {{{ source(system.file("OSGeo4W_test", package="rgdal"), echo=TRUE) }}} works, as well as PostGIS access, please let me know. ----------------------------------------------------- Installing FWTools is a convenient way to get started (but will be limited to the - extensive - range of drivers built into the binary package): url:[http://fwtools.maptools.org/] url:[http://fwtools.maptools.org/windows-main.html] Setting the FWTools bin/ directory in the PATH environmental variable is essential, (GDAL_DATA is set internally to the copy included in the rgdal package). The R Windows binary rgdal package can built against an FWTools Windows binary, using VC++. ...url[http://download.osgeo.org/fdo/3.4.1/binaries/] help for rgdal package[[BR]] url:[file:///C:/Program%20Files/R/R-2.10.0/library/rgdal/html/00Index.html][[BR]] [[BR]] Céline : "J'ai recherché dans mes bouquin et sur le Applied Spatial Data Analysis with R de Roger Bivand dont tu peux retrouver (une partie) des codes sur ce site" : url:[http://www.asdar-book.org/] Plus précisément sur cette page url:[http://www.asdar-book.org/book/die.R] ou à url:[http://www.asdar-book.org/book/die_mod.R]