Hi, Alex, I reorganized sortKeys slightly to allow the invocation of nosharing. Also added calls to nosharing in sortKeys and sortMerge. These calls should prevent the problem you are describing. Please let me know if you have an easy way for me to reproduce the problem. If you do have an suggestion, I would like to include it in the tests invoked by 'make check'.
Thanks. John On 5/20/11 5:07 AM, Alexandre Maurel wrote: > Hello John, > > While testing jNatural I found a problem which is probably related to > changes made to correct the sorting problem (ibis::util::sortKeys function), > you remove the calls : > > //vals.nosharing(); > //keys.nosharing(); > > In bord column the method ibis::bord::column::selectUInts return a > shallow copy of the internal buffer when the complete column is selected, > which cause jNatural count function to change the order of the > underlying column ... > > Regards, > > > _______________________________________________ > FastBit-users mailing list > [email protected] > https://hpcrdm.lbl.gov/cgi-bin/mailman/listinfo/fastbit-users _______________________________________________ FastBit-users mailing list [email protected] https://hpcrdm.lbl.gov/cgi-bin/mailman/listinfo/fastbit-users
