Hi, Roven,

If your columns always have the same number of subfields, I strongly
suggest that you break each subfield as its own column.

If your columns have variable number of subfields, they might be
considered as "set-valued" columns.  In this case, you may use the
ibis::keyword index.  You can take a look at file src/ikeywords.h for
an explanation of how.

Good luck.

John




On 7/5/13 1:40 AM, Roven Rommel Fuentes (IRRI) wrote:
> Hi Jerry,
> 
> I'm now studying a more complicated format that might require indexing.
> It is the vcf format for SNP data(same as the simple matrix I loaded
> in HDF5)  
> but each point is composed of more than one value. 
> For example: 
> 
> *X|Y*:val1:val2:val3:val4 (format)
> 
>            col1                       col2                        col3
> row1   G|G:48:1:51,51     A|G:48:8:51,51        A/A:43:5
> row2   T|T:49:3:58,50      T|A:3:5:65,3            T/T:41:3
> row3   G|T:21:6:23,27      T|G:2:0:18,2           T/T:35:4
>  row4   T|T:54:7:56,60      T|T:48:4:51,51       T/T:61:2
> 
> I need to get the X of row1 with val1>10 and val2<30.
> How should I implement this with FastQuery? All
> query is by row. If I'm not mistaken, FastQuery hasn't
> integrated yet the conditional query of FastBit but is there
> any way I can use index to do it? Should I just query an indexed
> row based on a variable(e.g var1) then filter offset before
> fetching Xs? Do you have other implementation in mind?
> 
> Thank you!
> 
> 
> Roven
> 
> 
>  
> 
_______________________________________________
FastBit-users mailing list
[email protected]
https://hpcrdm.lbl.gov/cgi-bin/mailman/listinfo/fastbit-users

Reply via email to