Question

problem loading lumi package, Error : object ‘sqliteQuickSQL’ is not exported by 'namespace:RSQLite'

0

Entering edit mode

lucashusquin ▴ 10

@lucashusquin-6959

Last seen 9.5 years ago

France

Hi All,

My name is Lucas, I'm currently doing an internship for my Master degree using Bioconductor to analyse the outcome of the HumanMethylation450 BeadChip on my cohort. I'm running ubuntu 14.04 LTS.

> sessionInfo()

R version 3.0.2

Platform: x86_64-pc-linux-gnu (64-bit)

Since yesterday (when I realized the issue, but it could have started earlier than that) I have not been able to load the lumi package. I had been using it for the past month and a half without any problems, but it won't work anymore.

When trying to load lumi I get the following error :

"Error : object ‘sqliteQuickSQL’ is not exported by 'namespace:RSQLite'"

I removed the lumi package from my library and tried to reinstall it using

> source("http://bioconductor.org/biocLite.R")

> biocLite("lumi")

The installation failed after downloading the packages and returned the following :

"Error : object ‘sqliteQuickSQL’ is not exported by 'namespace:RSQLite'
ERROR: lazy loading failed for package ‘lumi’"

According to some threads on different websites (http://seqanswers.com/forums/showthread.php?t=47785 particularly), RSQLite was updated to version 1.0.0 on October 25th, 2014 and this version of the package no longer contains the function "sqliteQuickSQL".

So I removed the current version of RSQLite that I must have updated since October 25th, and reinstalled the previous version (0.11.4), and then tried to reinstall lumi again.

After succesfully reinstalling lumi, under the 0.11.4 RSQLite version, I tried to load it

> library(lumi)

But I ended up with the error I had at the beginning :

"Error : object ‘sqliteQuickSQL’ is not exported by 'namespace:RSQLite'"

I do not know what to do to make it work, and I would greatly appreciate any help that you could provide.

Thanks in advance,

Lucas

lumi • 10k views

ADD COMMENT • link updated 9.0 years ago by laseaman • 0 • written 9.5 years ago by lucashusquin ▴ 10

0

Entering edit mode

Hi

I have the same problem, but I'm wanting to use 'CummeRbund', the main problem I have when trying to install the following packages Gviz, GenomicFeatures etc, is

Error : object ‘sqliteQuickSQL’ is not exported by 'namespace:RSQLite'

based on the answers below the problem is the version of R, but I can't get the latest. I'm using Fedora 18, I have uninstalled, reinstalled, updated R (yum update R, etc) but I can only get verison 3.0.2-1.fc18. Does this mean I have to upgrade my operating system to get Bioconductor to work?

As described below I have updated, reinstalled Bioconductor

> biocLite()
BioC_mirror: http://bioconductor.org
Using Bioconductor version 2.13 (BiocInstaller 1.12.1), R version 3.0.2.

Is there a way to install the latest version or R without using Fedora repos??

thanks in advance and sorry if this is a stupid question but this is the third night I've been struggling with this

Steven

ADD REPLY • link 9.5 years ago steven.yates • 0

0

Entering edit mode

You can always build the latest version of R yourself without waiting for it to propagate to the yum servers (note that there may be a way to get updated versions of R from yum, but not having used Fedora in about 10 years, I am not sure how one might go about doing such a thing).

Anyway, the way I usually install R (on Ubuntu, where I am not admin) is to download the sources from CRAN, untar in my home dir, and then build and make in place:

wget http://cran.r-project.org/src/base/R-3/R-3.1.2.tar.gz
tar xvfz R-3.1.2.tar.gz
cd R-3.1.2
./configure
make

And if you have sudo privileges, you can also install to your system-wide directories using

sudo make install

Or you could just add the ~/R-3.2.1/bin directory to your PATH (at the beginning of the path), or you could do

sudo ln -s /usr/bin/R-3.1.2 /path/to/home/R-3.1.2/bin/R

which is nice if you want/need to keep different versions of R laying about.

ADD REPLY • link updated 9.5 years ago by Dan Tenenbaum ★ 8.2k • written 9.5 years ago by James W. MacDonald 65k

0

Entering edit mode

HI Dan

thanks for the advice. I downloaded the tar.gz, unpacked and hit make but I had a slight problem it wouldn't make and the source of the problem was

error: --with-readline=yes (default) and headers/libs are not available

but this was fixed using

yum install readline-devel

Then everything worked and I have now installed Biocondctor and CummeRbund and everything is working now (fingers crossed). I thought I'd mention the above incase anyone else reads this and has the same problem, with Fedora

Once again many thanks Steven

ADD REPLY • link 9.5 years ago steven.yates • 0

0

Entering edit mode

Glad to hear it! Jim is the one you can thank. ;-)

ADD REPLY • link 9.5 years ago Dan Tenenbaum ★ 8.2k

score 1 · Answer 1 · 2014-10-31

You should update to Bioconductor 3.0 / R-3.1.2. Changes for compatibility with RSQLite_1.0.0 have only been made in Bioconductor 3.0 and the devel version (3.1).

So install R-3.1.2 and then do this:

source("http://bioconductor.org/biocLite.R")

If you see a message that says "Bioconductor version 2.14", do this:

biocLite("BiocUpgrade")

Then try installing lumi:

biocLite("lumi")

Dan

score 1 · Answer 2 · 2014-10-31

Thanks for your help,

I followed your steps and updated to a newer R-version, but I still have the error while trying to install the lumi package.

"Error : l'objet ‘sqliteQuickSQL’ n'est pas exporté par 'namespace:RSQLite'"

> sessionInfo()

R version 3.1.1 (2014-07-10)
Platform: x86_64-pc-linux-gnu (64-bit)

locale:
[1] LC_CTYPE=fr_FR.UTF-8       LC_NUMERIC=C
[3] LC_TIME=fr_FR.UTF-8        LC_COLLATE=fr_FR.UTF-8
[5] LC_MONETARY=fr_FR.UTF-8    LC_MESSAGES=fr_FR.UTF-8
[7] LC_PAPER=fr_FR.UTF-8       LC_NAME=C
[9] LC_ADDRESS=C               LC_TELEPHONE=C
[11] LC_MEASUREMENT=fr_FR.UTF-8 LC_IDENTIFICATION=C

attached base packages:
[1] parallel stats graphics grDevices utils datasets methods
[8] base

other attached packages:
[1] Biobase_2.22.0 BiocGenerics_0.8.0 RSQLite_1.0.0
[4] DBI_0.3.1 BiocInstaller_1.16.0

loaded via a namespace (and not attached):
[1] affy_1.40.0           affyio_1.30.0         annotate_1.40.1
[4] AnnotationDbi_1.24.0 base64_1.1            beanplot_1.2
[7] Biostrings_2.30.1     bumphunter_1.2.0      codetools_0.2-8
[10] colorspace_1.2-4      digest_0.6.4          doRNG_1.6
[13] foreach_1.4.2         genefilter_1.44.0     GenomicRanges_1.14.4
[16] grid_3.1.1            illuminaio_0.4.0      IRanges_1.20.7
[19] iterators_1.0.7       itertools_0.1-3       KernSmooth_2.23-10
[22] lattice_0.20-24       limma_3.18.13         locfit_1.5-9.1
[25] MASS_7.3-29           Matrix_1.1-2          matrixStats_0.10.3
[28] mclust_4.4            methylumi_2.8.0       mgcv_1.7-28
[31] minfi_1.8.9           multtest_2.18.0       nleqslv_2.5
[34] nlme_3.1-113          nor1mix_1.2-0         pkgmaker_0.22
[37] plyr_1.8.1            preprocessCore_1.24.0 RColorBrewer_1.0-5
[40] Rcpp_0.11.3           registry_0.2          reshape_0.8.5
[43] R.methodsS3_1.6.1     rngtools_1.2.4        siggenes_1.36.0
[46] splines_3.1.1         stats4_3.1.1          stringr_0.6.2
[49] survival_2.37-7       tcltk_3.1.1           tools_3.1.1
[52] XML_3.98-1.1          xtable_1.7-4          XVector_0.2.0
[55] zlibbioc_1.8.0

Do you have any other suggestions ?

Thanks again for your help

Lucas

score 0 · Answer 3 · 2014-10-31

0

Entering edit mode

James W. MacDonald 65k

@james-w-macdonald-5106

Last seen 4 hours ago

United States

It looks like you have a mix-n-match Bioconductor installation. You shouldn't have an RSQLite version anywhere near 1.0.0 with that old an R version. At this point I wouldn't even bother trying to fix your current R, but instead you should install R-3.1.1 and install Bioconductor from scratch using biocLite().

If you really want to use an old R-version, you can do

library(BiocInstaller)
biocValid()
biocLite()

Which will tell you which packages are off (biocValid()), and should install the correct ones (biocLite()). For those packages that are too new, biocValid() will tell you how to downgrade as well.

ADD COMMENT • link 9.5 years ago James W. MacDonald 65k

0

Entering edit mode

Jim, RSQLite is a CRAN package, so all versions of BioC are equally hosed by 1.0.0 (except 3.0 and 3.1 which have been fixed to deal with it).

ADD REPLY • link 9.5 years ago Dan Tenenbaum ★ 8.2k

0

Entering edit mode

Thanks for the pointer, Dan.

That's really unfortunate. I was under the mistaken impression that biocLite() did some version checking for CRAN packages as well. And note I am not saying anything, I'm just sayin'.

The reason this is unfortunate is because I (and possibly others) am often asked to do additional things for existing analyses. I try to follow the tenets of reproducible research by using Sweave files and dropping a sessionInfo() at the end so I can see what version I used. So when I need to add something to an analysis I did 2 years ago (yes, this happens, and yes, way more often than one would imagine...), I almost always add the extra bits to the Sweave document and re-run with whatever version of R/BioC I originally used. So I have lots of old R/BioC versions that I use semi-regularly.

And sometimes I have to install an additional package, which if it comes from BioC will be the correct package for that version of R/BioC, but if it comes from CRAN will most assuredly not be, particularly in the case of RSQLite.

So all future versions of me will have to remember to say 'n' when biocLite() helpfully asks if I want to update all these other packages when installing a package to an old install of R/BioC.

ADD REPLY • link 9.5 years ago James W. MacDonald 65k

0

Entering edit mode

Jim, we've discussed hosting a "frozen" CRAN repos that would be intended for use with a specific version of BioC.

I did make a snapshot of CRAN (pre-RSQLite_1.0.0) but haven't deployed it yet.

We did decide that we probably would not make these snapshots for no-longer-current versions of BioC, but that we'll do them going forward.

ADD REPLY • link 9.5 years ago Dan Tenenbaum ★ 8.2k

0

Entering edit mode

Nice! I can see that making snapshots of long-since-gone versions of BioC isn't worth the trouble, but it will be nice to have the snapshots going forward.

ADD REPLY • link 9.5 years ago James W. MacDonald 65k

score 0 · Answer 4 · 2014-10-31

0

Entering edit mode

Dan Tenenbaum ★ 8.2k

@dan-tenenbaum-4256

Last seen 3.2 years ago

United States

Update all your packages with

biocLite()

And answer "a" when prompted to update all outdated packages. That should fix it.

Dan

ADD COMMENT • link 9.5 years ago Dan Tenenbaum ★ 8.2k

score 0 · Answer 5 · 2014-10-31

0

Entering edit mode

lucashusquin ▴ 10

@lucashusquin-6959

Last seen 9.5 years ago

France

Dan, using biocLite() returns an error :

> biocLite()

BioC_mirror: http://bioconductor.org
Using Bioconductor version 3.0 (BiocInstaller 1.16.0), R version 3.1.1.
Message d'avis :
installed directory not writable, cannot update packages 'affy', 'affyio',
'annotate', 'AnnotationDbi', 'AnnotationForge', 'Biobase', 'BiocGenerics',
'BiocInstaller', 'bioDist', 'biomaRt', 'Biostrings', 'boot', 'BSgenome',
'BSgenome.Hsapiens.UCSC.hg19', 'bumphunter', 'class', 'cluster', 'codetools',
'FDb.InfiniumMethylation.hg19', 'foreign', 'genefilter', 'GenomicFeatures',
'GenomicRanges', 'illuminaio', 'IRanges', 'KernSmooth', 'lattice', 'limma',
'lumi', 'MASS', 'Matrix', 'methylumi', 'mgcv', 'minfi', 'minfiData',
'multtest', 'nlme', 'nnet', 'org.Hs.eg.db', 'preprocessCore', 'rpart',
'Rsamtools', 'RSQLite', 'rtracklayer', 'siggenes', 'spatial', 'sva',
'XVector', 'zlibbioc'

ADD COMMENT • link 9.5 years ago lucashusquin ▴ 10

0

Entering edit mode

Do you have root access on the machine? If so, run R as root:

sudo R

and try the same command.

ADD REPLY • link 9.5 years ago Dan Tenenbaum ★ 8.2k

0

Entering edit mode

It worked and all the packages were updated, but I still couldn't load the lumi package afterwards (same error message).

Do you have any other ideas ? It's 7:30pm here though so I will look at your answers over the week end, thank you again for your help and your time.

Lucas

ADD REPLY • link 9.5 years ago lucashusquin ▴ 10

0

Entering edit mode

Please post your sessionInfo() again after the latest attempt to load lumi.

Also, you seem to have a lot of packages loaded, try starting R as follows:

R --vanilla

and trying to install lumi as follows:

library(BiocInstaller)

biocLite("lumi")

ADD REPLY • link 9.5 years ago Dan Tenenbaum ★ 8.2k

0

Entering edit mode

Hello Dan,

Thanks a lot it worked this time !

Lucas

ADD REPLY • link 9.5 years ago lucashusquin ▴ 10

score 0 · Answer 6 · 2015-05-07

I was having this same problem (the cummeRbund version) and because I was working on a server I did not have the permissions to update the packages. I found a workaround though. If you just run these two lines, before running the readCufflinks command after library(cummeRbund) it should work. I think it would likely help with the original lumi error but I wasn't having that problem so I don't know for sure.

sqliteQuickSQL<-dbGetQuery
dbBeginTransaction<-dbBegin

found here: http://aferrarini.altervista.org/pages/cummeRbund_workaround.html