I am getting started to learn tiledb but am now stuck on how I can lay out the data. I have time series data where I have several data points at each time point.
I get these values in column major order:
1000,333,23,2,-34, 1010,334,33.2,-134, …
and do not want to reorder these values if possible somehow.
How should I choose dimensions and attributes to have good write performance and later good read performance when slicing by time range and read mostly all values?
Would sorting the values by type help?
Thanks for any help,