Consider a table of ohlcv data:

date       ticker    open   close  high   low  volume exch       time                         
-----------------------------------------------------------------------------------------------
2024.05.02 NVDA 173    173.7  173.95 173    870      NASDAQ      2024.05.02D09:15:00.000000000
2024.05.02 NVDA 173.85 173    173.85 173    2290     NASDAQ      2024.05.02D09:16:00.000000000
2024.05.02 NVDA 173.55 173.5  173.55 173.25 382      NASDAQ      2024.05.02D09:17:00.000000000
2024.05.02 NVDA 173.5  173.5  173.5  173.5  10       NASDAQ      2024.05.02D09:18:00.000000000
2024.05.02 NVDA 173.65 173.05 173.65 173.05 435      NASDAQ      2024.05.02D09:19:00.000000000
2024.05.02 NVDA 173.1  172.9  173.1  172.9  1902     NASDAQ      2024.05.02D09:20:00.000000000
2024.05.02 NVDA 172.6  172.45 172.6  172.05 824      NASDAQ      2024.05.02D09:21:00.000000000

and it’s sorted on both ticker and time via

ohlcv: `ticker`time xasc ohlcv;

yielding

q)meta ohlcv
c     | t f a
------| -----
date  | d    
ticker| s   s
open  | f    
close | f    
high  | f    
low   | f    
volume| f 
exch  | s    
time  | p 

Clearly from the above, only ticker is sorted, but time is not. However, from the data one also observes that the table is both partitioned and grouped on the ticker column which leads me to my question.

Which of these attributes do I choose to set? And how about the time column? I suppose it’s application dependent. If so, what’s the best practice?