It's worth mentioning a few things for background context:
"PartiQL" is a specification for a SQL dialect capable of querying data from various sources (whether structured/relational or not):
> "PartiQL provides SQL-compatible query access across multiple data stores containing structured data, semistructured data, and nested data. It is widely used within Amazon and is now available as part of many AWS services, including DynamoDB."
Amazon has an internal implementation of the PartiQL query language specification, that is used to make DynamoDB (and some other sources) queryable with SQL:
This is similar in concept to something like Google's ZetaSQL, or Apache Calcite in a way.
So the library in the link is a tool to generate Kotlin code given a description of an IR/AST format in Ion language (another Amazon invention), along with visitors and tree-transformers/rewrite interfaces, etc.
The assumption I would make is that they are probably using/intend to use this to write the query engine powering the "PartiQL on DynamoDB" service (maybe I am wrong). So I'd imagine it's pretty solid code.
"PartiQL" is a specification for a SQL dialect capable of querying data from various sources (whether structured/relational or not):
Amazon has an internal implementation of the PartiQL query language specification, that is used to make DynamoDB (and some other sources) queryable with SQL:https://docs.aws.amazon.com/amazondynamodb/latest/developerg...
This is similar in concept to something like Google's ZetaSQL, or Apache Calcite in a way.
So the library in the link is a tool to generate Kotlin code given a description of an IR/AST format in Ion language (another Amazon invention), along with visitors and tree-transformers/rewrite interfaces, etc.
The assumption I would make is that they are probably using/intend to use this to write the query engine powering the "PartiQL on DynamoDB" service (maybe I am wrong). So I'd imagine it's pretty solid code.
It certainly looks like it from the README.