SaarDB Dev Log: Bridging SQL Parsing and Key-Value Storage with Binary Encoding

In the sixth installment of the SaarDB development series, the author tackles how parsed SQL statements are translated into key-value operations compatible with the database's storage engine. The storage engine has no native understanding of tables, columns, or data types, so a bridging layer is needed to map SQL constructs like CREATE TABLE and INSERT INTO PUT operations. Table schemas are stored using a reserved key prefix format, such as '_schema:<table_name>', with the schema data as the value. Rather than using JSON or full struct serialization — which are space-heavy and compute-intensive — the author adopts a compact binary serialization strategy where field order is agreed upon in advance. This approach encodes fixed-length integers directly and prefixes variable-length strings with their length, consistent with the binary format used in earlier parts of the series.