When it comes to modeling over 1,000,000 data points in JSON, there are two potential paths to consider:
a) Utilizing an array of objects:
[{time:123456789,value:1432423},{time:123456790,value:1432424},....]
or
b) Implementing nested arrays
[[123456789,1432423],[123456790,1432424],....]
On initial comparison, the latter option may seem faster due to its minimized character usage, although it sacrifices some descriptive detail. Is the speed difference significant between these two methods? Which one would be your preference and for what reasons?
Could there potentially be a third approach worth exploring?