[Rd] RSQLite indexing: summary

Thomas Lumley tlumley at u.washington.edu
Wed Oct 24 01:09:23 CEST 2007


I asked about slow indexing in RSQLite for a genetic database.  Seth Falcon's suggestion of making sure that the identifiers were stored as integer rather than string made a big difference.  SNPs come from the factory as "rs100092" and stripping the "rs" off the front is easy.

Other advice about larger or smaller SQLite cache size didn't seem to have much impact in my setting, and I didn't try the advice about getting a different database.

Despite it's many other virtues, SQLite is still slow at indexing.

Thanks to all.

     -thomas=

Thomas Lumley			Assoc. Professor, Biostatistics
tlumley at u.washington.edu	University of Washington, Seattle



More information about the R-devel mailing list