One thing I learned about: Bloom filters.
On-chain data indexers often try to avoid refetching the same transactions.
A bloom filter can help to test if they already have a record.
- It is a probabilistic data structure consisting of a binary array
- Maps record hash values to values of the array
- Maps record keys to the corresponding indexes of the array
- Membership check consists of checking the array values
Caveat: Can produce false positives.
🔽🛠️Resources🔽
- Real-time Blockchain Data Indexing: Challenges and Solutions (2025): https://medium.com/@aannkkiittaa/real-time-blockchain-data-indexing-challenges-and-solutions-fcb1f8aa3911
- Bloom Filters (2022): https://www.youtube.com/watch?v=V3pzxngeLqw&t=208
Top comments (0)