> From: Alaric Snell-Pym <xxxxxx@snell-pym.org.uk>
> Date: Tuesday, October 01, 2019 6:09 AM
>
> On 01/10/2019 10:54, xxxxxx@ancell-ent.com wrote:
>
> [ JDBC to be preferred over MariaDB CONNECT ]
>
>> Looking at them one way, both JDBC and MariaDB CONNECT are ways of
>> allowing sdbi to access more "databases" than the ones we'll be
>> (eventually) supplying more direct methods for.  Scare quotes if
>> you don't consider a CSV file to be a database.  Users can choose
>> either if both for example support a serious database like Db2,
>> but JDBC will be suggested option.
>
> To be precise, I'd consider a CSV file perfectly eligible to be part
> of a database. Some people have tried to draw some lines in the sand
> and define "what is a database" - eg, ACID properties - but this
> usually just encodes their prejudices, and inspires others to find
> useful ways to store data that sits outside of that definition...

Per John's insight, all we really care about is if a "database"
provides rectangular results, so "databases" like CSV that start out
being rectangular are a very natural fit.

ACID is definitely not required, we in theory support any NoSQL/BASE
(Basic Availability, Soft State, Eventual Consistency,
https://www.lifewire.com/abandoning-acid-in-favor-of-base-1019674)
database if has a text query language and supplies rectangular results.
For CSV files and other not a formal database file formats, MariaDB
CONNECT provides both.

Since SQL is so ubiquitous, NoSQL databases like Apache Cassandra (wide
column, so I think this is pretty natural) and and Neo4j (a bit of a
coercion) also have SQL style query languages which return rectangular
results.  And per the latter, ISO/IEC is standardizing a Graph Query
Language (GQL) in part inspired by their's, at the same level as SQL:
https://neo4j.com/blog/gql-standard-query-language-property-graphs/

- Harold