Hi, Thiago, I presume that the ith row of X is to be considered part of the ith row of the whole dataset/data table along with ith row of Y and Z. If this is the case, you need put all three of them in a single data partition (i.e., in the same directory). There is any tool to help you merge the data at this point, we should put this as an item on our to-do list..
John On 10/21/2009 5:42 AM, Thiago L L Siqueira wrote: > Hi! > > I need some assistance for the following problem: > > I want to use the Bitmap index for attributes X, Y and Z. > The index for attribute X shoud be "<binning none/><encoding equality/>". > The indices for attributes Y and Z should be "<binning nbins=1000, > equal-weight/><encoding equality/>". > > I have already: > > 0) generated a csv file for attribute X, and a csv file for attributes > Y and Z (both have the same number of rows) > 1) created a partition for attribute X in the directory x > 2) created a <binning none/><encoding equality/> for partition x > 3) created a partition for attributes Y and Z in the directory yz > 4) created the indices <binning nbins=1000, equal-weight/><encoding > equality/> for partition yz > 5) queried using the following ibis command: ibis -d c:\x c:\yz -q > select X where Y>=y1 and Y<=y2 and Z>=z1 and Z<=z2" > > However, the cited ibis command issued messages similar to "Y>=y1 and > Y<=y2 and Z>=z1 and Z<=z2 cannot be processed on data partition x". > The query result has 0 rows, which is not true. Clearly, partition x > does not contain Y and Z attributes, but how can I solve this query? > As far as I know, a table can be stored in multiple partitions. On the > other hand, I need to use different kinds of indices for each > attribute. In fact, attribute X should not have binning. Is it > possible? > > Thanks in advance, > _______________________________________________ FastBit-users mailing list [email protected] https://hpcrdm.lbl.gov/cgi-bin/mailman/listinfo/fastbit-users
