You'll only be able to do analytics on current data, or if you want historical data you'd need to start recording it and can only do historical data from that point on, unless you were to take past historical data from offchain and then save it to the smart contract.
All of this is a moot point though, because it would be prohibitively expensive and very unrealistic to do. Not to mention you'd probably run out of block space.
Ok thanks. Noob question. By doing this, how do you run out of block space?
It'd likely be a very large transaction and depending on how large of a dataset you're working with you can potentially fill up an entire block with that one transaction. Granted, in reality a transaction that size will probably never make it into a block unless you pay a good validator tip since it'd be more profitable to leave that transaction out and get MEV from a bunch of smaller transactions.
Ok thanks for your input.
why would you want that? It would cost money, a smart contract doesn't have wide visibility, no storage, ...
or maybe I missed the point
If the data analytics (along with storing the data) can be done on-chain in the smart contract itself, then this may perhaps reduce any latencies when the data analytics (along with storing the data) is done off-chain somewhere else (which relies on a different system which may perhaps have latencies).
Yeah this would cost money since a smart contract needs to be deployed.
A smart contract does have wide visibility though because it’s a smart contract deployed on the blockchain, hence it is transparent to everyone.