Hi, Petr, You are on the right track. Unfortunately, there isn't anything currently to support what you've suggested. A relatively straightforward version of this could become available soon - it is one of the things I am working on now. An optimistic estimate is that it would take another three weeks. I will check the code in as soon as I am comfortable that it works.
John On 11/7/11 4:59 AM, Thorgrin wrote: > Hello John, > > since aggregation is somewhat greedy process in fastbit, it occurred > to me that if I run the query for each partition separately and then > combined the result (and aggregated again), I could get the result > with less memory. Is it possible to do something like this? > > I cannot find any function that would allow me to combine multiple > ibis::table classes, only ibis::part, which is not enough in this case. > > This would not be an issue if there was some way to aggregate more > efficiently. > > What do you think? > > Petr Velan > > > _______________________________________________ > 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
