graph-node VS log-flume

A subgraph according to The Graph (which is a decentralized protocol for indexing and querying blockchain data) is a custom API built on blockchain data. They are queried using the GraphQL query language and are deployed to a Graph Node using the Graph CLI.

Event is a log entity which EVM smart contracts can emit during transaction execution. Events are very good at signalling that an some action has taken place on-chain. Applications can subscribe and listen to events to trigger some off-chain logic or they can index, transform and store events in some off-chain storage (look at The Graph protocol or Ethereum ETL).

Subgraph cloned from repo :https://github.com/graphprotocol/graph-node/tree/master/docker

https://github.com/scaffold-eth/scaffold-eth-2 this repo comes with wagmi and hard hat preconfigured you can use it as a sample , but if you are large querying data I would advise you the graph https://thegraph.com/

The Graph

There are also decentralized network services like The Graph or Pocket Network that provide decentralized access to Ethereum and other blockchain networks. Note that these services primarily provide APIs for querying blockchain data, and may not provide full functionality for sending transactions or executing smart contracts like a full Ethereum node or services like Infura and Alchemy.

For read-only the way to use The Graph

See it for yourself.

thegraph.com is the only site operated by The Graph Foundation, and is the centrepiece for all activity on the protocol. Can a website/web dApp be hacked? Yes. Has it been hacked in the past two years since it launched? No.

Check out TheGraph, most of DEX has a subgraph indexing data

Excluding State data, the team at Rivet.cloud has built Flume, which indexes everything into a Sqlite database. Flume offers a number of baked in RPC queries, but you can also open the sqlite database and run your own queries. I should caution you that for ETH mainnet, the database is about 1.6 TB.

I'd also point you to Flume, a project that already does this. I work for Rivet.cloud, and we use Flume to index blocks, transactions, receipts, and logs for faster querying. It can sync up over Websockets (very time consuming) or you can give it a synced Geth database and it can index it much more quickly.

Block explorers like Etherscan and EthVM maintain additional indexes that use a lot more space than a normal node to make everything efficiently queryable. My team has developed Flume which will index logs, transactions, receipts, and blocks into a relational database, but that database ends up being >1.25 TB without even having the Ethereum state trie. Queries are super efficient, but the tradeoff is a lot more disk space.

We also have an open source log and transaction indexer called Flume, which gives high performance responses to eth_getLogs queries, and supports more in-depth queries of log and transaction data than are supported through traditional Ethereum nodes.

Flume will load blocks, transactions, receipts, and logs into a sqlite database, from which it has a number of rpc queries, though locally you could connect to the sqlite database and run your own queries. It doesn't include state data, however.