partiql-lang-kotlin VS cbor-java

Ion is heavily used on the retail side of Amazon, but it's only recently started to appear in AWS products.

AWS is starting support PartiQL (https://partiql.org/) queries in some places and PartiQL uses Ion's type system internally.

Yep, that's exactly the vision here. One of the biggest sources of precedent and inspiration for us here is https://partiql.org/ which is picking up a fair bit of traction across AWS. Also see https://rockset.com/

QLDB allows you to create a ledger that acts similar to a schema or table space in a traditional database. Once you've created this ledger, you create a table through the SQL-like query language PartiQL that also enables you to interact with the data.

PartiQL[0] is an open source library that is a superset of SQL that I really like. It supports querying nested structures inside columns, so if a column contains some JSON data you can use the standard dot notation to query nested JSON data directly

[0] https://partiql.org/

PartiQL was introduced to AWS DynamoDB, with AWS making the announcement in 2020 making the life of developers easier, with the comfort of executing commands similar to SQL.

I haven't tried it, but I'm pretty sure the problem is that timestamp (and it's various capitalizations) is a keyword in PartiQL (https://github.com/partiql/partiql-lang-kotlin/blob/master/lang/src/org/partiql/lang/syntax/LexerConstants.kt#L221). To get it to be interpreted as an attribute name, you need to enclose TimeStamp in double quotes.

What's the pros and cons of this versus CBOR, which we had great success with in our system.

https://cbor.io

CBOR (Concise Binary Object Representation) is a codec for small JSON object, similar to MessagePack but in the internet standard.

JSON is a simple key=value pair representation. If you want to implement it by your hand with the same set of features (maps, lists, character escaping) then you will get something really similar. They are binary alternatives like https://cbor.io/ , but they are not popular for same reasons we use human readable representation of data for any other use case in programming

CBOR

Maybe this? http://cbor.io/ Haven't used it personally but I've read good things about it and the documentation makes sense.

go-cose is the Go implementation of COSE. COSE is a signing and encryption format based on the Concise Binary Object Representation format (CBOR). While inspired by the needs of the IoT community including fast processing on low-memory devices, the format is broadly applicable from small devices to large-scale server environments. COSE has incorporated lessons learned from Javascript Object Signing and Encryption (JOSE) efforts including JSON Web Signature (JWS) and JSON Web Token (JWT).

Have a look at cbor, https://cbor.io/ It's very compact and works wonderfully for embedded devices. Works identical to Json but than for binary data.

You could also look into CBOR serialization format. It's less readable but it's much more efficient than JSON.