Guys,
I tried to use "GenomicFeatures" to create trxDb. But I kept hitting the same error as below. Actually, I have plenty of disk space and I really don't know how to solve this issue. Thank you for your any help!
hse<-makeTxDbFromGFF("/genome/human/hg19_ensemble.gtf",format="gtf")
Prepare the 'metadata' data frame ... metadata: OK
Error in sqliteSendQuery(con, statement, bind.data) :
rsqlite_query_send: could not execute1: database or disk is full
> sessionInfo()
R version 3.2.1 (2015-06-18)
Platform: x86_64-unknown-linux-gnu (64-bit)
Running under: Red Hat Enterprise Linux Server release 6.6 (Santiago)
locale:
[1] C
attached base packages:
[1] stats4 parallel stats graphics grDevices utils datasets
[8] methods base
other attached packages:
[1] GenomicFeatures_1.20.1 AnnotationDbi_1.30.1 Biobase_2.28.0
[4] GenomicRanges_1.20.5 GenomeInfoDb_1.4.1 IRanges_2.2.5
[7] S4Vectors_0.6.1 BiocGenerics_0.14.0
loaded via a namespace (and not attached):
[1] XML_3.98-1.3 Rsamtools_1.20.4 Biostrings_2.36.1
[4] GenomicAlignments_1.4.1 bitops_1.0-6 futile.options_1.0.0
[7] DBI_0.3.1 RSQLite_1.0.0 zlibbioc_1.14.0
[10] XVector_0.8.0 futile.logger_1.4.1 lambda.r_1.1.7
[13] BiocParallel_1.2.7 tools_3.2.1 biomaRt_2.24.0
[16] RCurl_1.95-4.7 rtracklayer_1.28.6
Hi,
Could it be that you are running out of RAM?
makeTxDbFromGFF()and all the functions in themakeTxDb*family create the SQLite db in memory, not on disk. Note that the error message you get, which is coming from the SQLite C code itself and relayed by the RSQLite package, is misleading.H.
Herve,
Thank you for your answer and I am sorry that I cannot reply to you earlier. Actually, I don't think it's a memory issue as our server on cluster has a lot of memory. Also, I have to do this step to create trxDB using my laptop since I cannot do it using the server. It works fine without any problem.
So, I am wondering if the server misses some parts of components which lead to broken SQLite db issue?
-X
Are there complicated file permissions differences between the R process you're running under and the file system, or perhaps the file system is not local and involves some implicit authentication step? Maybe writing to /tmp would work?