That really helps.

Setting up the new data source isn't a problem.

The B part:

JSON import/export - that looks like the easiest way to go. In each data point, I'd just have to change dataSourceXid and the settings under pointLocator to reflect JSON fields instead of RegEx fields.

Sounds not too hard, but I find documentation is a bit lacking on the NoSQL tools.
It would be useful is mango provided an easy point history import/export feature. I know in the watch-list I can export CSV, but it would be nice to have an "export all" for a datapoint somewhere that could be easily imported back in. Like an additional tab to import/export data points.

Sounds more complicated than the route I'd like to go down. Keeping the XID the same wouldn't break things elsewhere.

I don't know how often it happens, but maybe a future tool to move a data point to a new data source would be useful?

@CraigWeb That's a good re-assurance I'm on the right train of though.

It would be neat if there was a data point that could capture sunrise and sunset for a given location. Changes in those values would trigger a context change and update the statistical point. Sunrise and Sunset values would be great if they could be used in the advanced scheduler, too, as one could save energy by using those values to trigger events like setpoints and the like.

I was hoping to use the envcands Environment Canada module to capture weather data and perhaps sunrise/sunset, but I never managed to make it work. There's a distinct lack of documentation on that module. If there is documentation, that would be extremely useful.

@tomatopi

This code is assigning the point-value object to a variable in the local scope called water. Since we are working in javascript, variables must be valid javascript variables.

@joeamiraglia That is exactly what I have running now on a test point and it does work.

@MattFox Your solution would work, but as @CraigWeb said, charting and displaying the data proved to be a little more difficult.

I'm hoping to avoid re-programming my sensor units to pre-parse the data (I'm using custom make microcontrollers). That would make them less flexible if I add/remove/change sensors.

I originally assumed I'd just need a parsed value to determine if it's moist or dry. Since then, I noticed the history chart was very useful. Since we're tracking soil moisture in a pot, displaying the values in a chart compared to sunlight, temperature, humidity, etc. is proving a lot more useful that I expected.

It would be nice if the modbus datapoint (or others) were able to call a global script instead of just the multiplier/added function. That would be the bees knees.

I'm not entirely sure the direction I will go, but I'm leaning to biting the bullet and using the two points per solution just to get the future flexibility to chart and trigger events.