On 02/24/2011 05:20 AM, Dan Tenenbaum wrote: > On Thu, Feb 24, 2011 at 1:58 AM, Aszodi,Andras <andras.aszodi at="" imp.ac.at="">wrote: > >> >> On Feb 23, 2011, at 8:09 PM, Dan Tenenbaum wrote: >> >> >> >> On Wed, Feb 23, 2011 at 3:07 AM, Aszodi,Andras <andras.aszodi at="" imp.ac.at="">wrote: >> >>> Dear Mailing List, >>> >>> When I say:- >>> >>>> library("RdbiPgSQL") wanted to mention RPostgreSQL on CRAN as an alternative. Martin >>> >>> ...then I get the following errors:- >>> >>> Loading required package: Rdbi >>> Error in dyn.load(file, DLLpath = DLLpath, ...) : >>> unable to load shared object >>> >>> '/Library/Frameworks/R.framework/Versions/2.12/Resources/library/R dbiPgSQL/libs/x86_64/RdbiPgSQL.so': >>> >>> >>> dlopen(/Library/Frameworks/R.framework/Versions/2.12/Resources/lib rary/RdbiPgSQL/libs/x86_64/RdbiPgSQL.so, >>> 6): Symbol not found: _PQbackendPID >>> Referenced from: >>> >>> /Library/Frameworks/R.framework/Versions/2.12/Resources/library/Rd biPgSQL/libs/x86_64/RdbiPgSQL.so >>> Expected in: dynamic lookup >>> >>> Error: package/namespace load failed for 'RdbiPgSQL' >>> >>> I use only binary packages for R and R itself was installed as binary. But >>> I >>> must admit my Postgres installation (8.4.7) was compiled from source and >>> placed >>> into a "nonstandard" location /usr/local/pgsql/current . But the problem >>> persists even if I put the library path /usr/local/pgsql/current/lib in >>> the >>> DYLD_LIBRARY_PATH variable. This happens on a Mac Pro (64-bit Intel), with >>> OS X >>> 10.6.6. I would be grateful for any fixes or workaround advice. Thank you >>> very >>> much! >>> >>> PS I first sent this to the R bug tracking system but Professor Ripley >>> advised me to submit it to Bioconductor since RdbiPgSQL is a BioC >>> contributed package. I hope this is now the right place... >>> >>> >> Hi, you don't include the output of sessionInfo(), >> >> >> Sorry, my bad. Here it is. I wonder if this helps though... >> >>> sessionInfo() >> R version 2.12.1 (2010-12-16) >> Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit) >> >> locale: >> [1] C >> >> attached base packages: >> [1] stats graphics grDevices utils datasets methods base >> >> other attached packages: >> [1] Rdbi_1.24.0 >> >> loaded via a namespace (and not attached): >> [1] tools_2.12.1 >> >> nor explain how you installed RdbiPgSQL in the first place >> >> >> I said "I use only binary packages for R", see above. So RdbiPgSQL was >> installed with bioClite as usual. Definitely not from source. >> >> and whether there were any errors. >> >> >> No, there were no errors. >> >> >> You might try building RdbiPgSQL from source: >> >> source("http://bioconductor.org/biocLite.R") >> biocLite("RdbiPgSQL",type="source") >> >> This will give you some informative messages about how to tell the build >> process about your PostgreSQL installation. >> >> >> This is a good suggestion. What bothers me though that even if I specify >> LD_LIBRARY_PATH or DYLD_LIBRARY_PATH it won't help. Are the shared library >> locations somehow hard-coded in the pre-build packages? >> >> Yes. > > >> >> The pre-built version expects PostgreSQL to be in a standard location. >> >> >> That might be the problem. However, for Postgres there is no unequivocal >> "standard location" under OS X. Some people put it in /usr/local, >> Linux-style. There are pre-built versions that put the whole thing under >> /Library, MacPorts uses /opt I believe... it is a mess. My reason for the >> not-quite-standard location is that I plan to run 8.4 and 9.0 as well. >> >> > See here for how it is built on our Mac build system: > http://bioconductor.org/checkResults/release/bioc-LATEST/RdbiPgSQL /pelham-checksrc.html > Note the occurrences of /usr/local/include and /usr/local/lib. > > When I add the 'type="source"' flag to biocLite("RdbiPgSQL"), I see the > following; > > I could not find your PostgreSQL client libraries! > Use --with-pgsql-libraries=PATH; if running R's INSTALL, > use --configure-args='--with-pgsql-libraries=PATH'; or > set PG_LIB_DIR in your environment to the library path, > and rerun the configure/install. > > Passing these arguments by adding them to configure.args in your call to > biocLite(), and/or setting environment variables as described, should > probably build the package successfully for you. > >> Dan >> >> Many thanks for your prompt reply and for your suggestions! >> Andr?s >> >> Dan > > [[alternative HTML version deleted]] > > > > > _______________________________________________ > Bioconductor mailing list > Bioconductor at r-project.org > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor -- Computational Biology Fred Hutchinson Cancer Research Center 1100 Fairview Ave. N. PO Box 19024 Seattle, WA 98109 Location: M1-B861 Telephone: 206 667-2793

On Feb 24, 2011, at 2:50 PM, Martin Morgan wrote: > wanted to mention RPostgreSQL on CRAN as an alternative. Martin ...and this is what I did in the end, based on the suggestions from http://blog.earlh.com/index.php/2009/12/querying-postgres-or- greenplum-from-r-on-a-mac/ . Detailed steps follow for the sake of posterity. Execute them all in a root shell. ${POSTGRES} below refers to the root of your Postgres installation that contains the include/, lib/, bin/, shared/ subdirectories. 1) Make sure the Postgres header files are found: export PG_INCDIR=${POSTGRES}/include/ 2) Make sure GCC can find the pg_config executable: export PG_CONFIG=${POSTGRES}/bin/pg_config 3) Invoke R. Make sure you invoke the 64-bit version if your Postgres installation is 64-bit (mutatis mutandis for 32-bit installations) >From within R, say install.packages('RPostgreSQL', type='source') select the CRAN mirror, and watch the package being compiled. It is actually quite easy once you know what to do... :-) Many thanks to Dan and Martin for their kind help and suggestions. best regards, Andr?s