> Not yet. I would first like to know what is the time consuming part and
> what is a work around. If you are sure individual columns for every
> entry of the array solve the issue I will joyfully implement it. The
> downsize of this approch is that the array dimensions are not always the
> same in my scenario. But I have a workaround in mind for this issue.
The first thing I would try would be to completely normalize te file, aka
create table data as
(
id int,
t timestamp,
map_x int,
map_y int,
value float
);
and go with denormalized approach only when this doesn't work for some reason.
merlin