On 10/4/21 12:36 PM, Israel Brewster wrote: [snip]
Indeed. Table per station as opposed to partitioning? The *most* I can reasonably envision needing is to query two stations, i.e. I could see potentially wanting to compare station a to some “baseline” station b. In general, though, the stations are independent, and it seems unlikely that we will need any multi-station queries. Perhaps query one station, then a second query for a second to display graphs for both side-by-side to look for correlations or something, but nothing like that has been suggested at the moment.
Postgresql partitions *are* tables. What if you partition by station (or range of stations)?
-- Angular momentum makes the world go 'round.