How ClickHouse Parses CSV in Parallel
ClickHouse splits CSV into row-aligned chunks with a quote-aware scanner, parses them on separate threads, and reorders them with a ticket ring.
Performance
How ClickHouse Resolves a Scalar Function Call
ClickHouse’s function factory: static registration, alias maps, and a declarative signature DSL for overload resolution.
How ClickHouse Runs UNION, INTERSECT and EXCEPT
UNION ALL merges streams, UNION DISTINCT adds a hash DISTINCT with a self-abandoning preliminary pass, and INTERSECT/EXCEPT filter the left side through a set of the right.
How ClickHouse Stores Dictionaries in Sharded Hash Tables
The hashed dictionary packs keys compactly, shards the table for parallel loading, and picks its hash table by measured memory footprint.
How ClickHouse Writes the Native Format, Column at a Time
The Native format writes a row count, then for each column its name, type, serialization kind and bulk data — never a row.