I was looking to see if there was more documentation on -part.txt for the Ibis 
implementation.

In particular, is it possible to use an binary data that is a appended arrays 
of columnar data, by providing the offset per array? As a related question, 
could you specify a struct offset if the data happen to be an array of structs 
(this particular use case isn't, but other data might be).

Or does FastBit fundamentally require that the data are written as 0 offset, 
dense columnar binary data?

If not, how easy/hard would it be to use the library to be able to generate 
indices on that sort of data? Essentially, the use case will not/cannot 
reformat/copy the data because it is too big. Though, it is essentially 
columnar, but not separate files per column (columns are offsets into a single 
file).

Thanks,
-Jon
_______________________________________________
FastBit-users mailing list
[email protected]
https://hpcrdm.lbl.gov/cgi-bin/mailman/listinfo/fastbit-users

Reply via email to