EVM-focused UDFs for ClickHouse – Accelerate Ethereum & EVM-compatible blockchain analytics

clickhouse-evm is a collection of high-performance user-defined functions (UDFs) that extend ClickHouse with capabilities tailored for Ethereum Virtual Machine (EVM) data processing.

Whether you're building blockchain explorers, indexing on-chain data, or running deep analytics on Ethereum or other EVM-compatible chains, this project brings native decoding, parsing, and querying support into your ClickHouse workflows.

• 🧠 Decode EVM calldata, logs and transactions directly within ClickHouse
  • ethereum_decode_tx
  • evm_decode_call
  • evm_decode_calldata
  • evm_decode_event
• 🔄 Keccak-256 hashing UDF for topic and selector lookups (and many more)
  • keccak256 (now avalaible natively in ClickHouse)
• 🧱 Utility functions for working with EVM 0x-prefixed hex encoding
  • evm_hex_encode
  • evm_hex_decode
  • evm_hex_encode_int
  • evm_hex_decode_int
• Utility functions for working with EVM ABI documents, event and function signatures
  • evm_descriptor_from_fullsig
  • evm_signature_from_descriptor
  • evm_hex_encode
  • evm_hex_encode
• ⚡ Fast, optimized RPC calls to EVM-compatible nodes directly from ClickHouse queries
  • ethereum_rpc
  • ethereum_rpc_call
• 🚀 Speeds up on-chain data analysis by reducing external parsing overhead

• Quickly extract function parameters from calldata in a ClickHouse query
• Filter and parse smart contract events by ABI signature
• Analyze token transfers, contract interactions, or DeFi protocol data at scale
• Query live on-chain data through efficient RPC calls from within ClickHouse
• Integrate with your existing ClickHouse-based blockchain indexing pipeline

The output of the build process is distributed as a compressed archive called a bundle. This bundle includes everything needed to deploy and use the UDFs in ClickHouse.

Each bundle contains:

• 🧩 Standalone binary implementing the native UDFs (compiled with ClickHouse compatibility)
• ⚙️ ClickHouse configuration files (.xml) to register each native UDF
• 📝 SQL files for SQL-based UDFs (used for lightweight functions where SQL outperforms compiled code)

make bundle              # Build for native execution
GOOS=linux make bundle   # Cross-compile for use in Docker (Linux target)

This will:

• Generate the bundle directory at tmp/bundle/
• Create a compressed archive at tmp/bundle.tar.gz

The internal file structure of the bundle reflects the default layout of a basic ClickHouse installation.
As a result, decompressing the archive at the root of a ClickHouse server filesystem should "just work" with no additional path configuration.

clickhouse local \
    --log-level=debug \
    --path tmp/clickhouse \
    -- \
    --user_scripts_path="./tmp/bundle/var/lib/clickhouse/user_scripts" \
    --user_defined_executable_functions_config="./tmp/bundle/etc/clickhouse-server/*_function.*ml" \
    --user_defined_path="./tmp/bundle/var/lib/clickhouse/user_defined"

This runs ClickHouse in local mode using the provided config and a temporary storage path.

docker compose up -d

This launches a ClickHouse server inside a Docker container using the configuration and UDFs from the bundle.

docker compose -f docker-compose.prod.yaml up -d