[BioC] matrix transformation
Ido M. Tamir
tamir at imp.ac.at
Mon Oct 18 12:50:02 CEST 2010
> Dear Mailing List,
> I wrote the following function "matrix2tuple.sf" to translate a "cartesian
> xy matrix" as a "tuple matrix", to store it in relational database. The
> code works fine for a test set. My problem is, my real matrix is 7000 x
> 10000 big, which ends up in 70'000'0000' tuples. Transformation takes days
> Has anyone an idea, how I can optimize the described functions for speed?
> Best Wishes, Elmar Bucher
I didnt look at the R code, but the inserts could be sped up if you would
do it in batches and not individual inserts.
100 x insert..
Maybe do a cheap profiling by just commenting out the db code to see
if the db is slowing it down.
More information about the Bioconductor