> From: Alaric Snell-Pym <xxxxxx@snell-pym.org.uk>
> Date: Tuesday, October 01, 2019 4:26 AM
>
> On 30/09/2019 16:25, Lassi Kortela wrote:
>
>>> That is not at all clear to me.  As I pointed out, both designs
>>> involve three processes: the client, the server, and an intermediate
>>> program involving Java.
>>
>> I expect the solution without MariaDB will be much lighter. DB
>> engines are quite big pieces of software and non-trivial to install.
>
> Let me get this straight (I'm catching up on these emails, having had
> not enough sleep): We're wondering how to connect to the rich world of
> JDBC and the options on the table are:
>
> 1) Use a Java subprocess to use the JDBC libraries directly, sending
> actual SQL query strings into the JDBC implementation that get sent to
> the actual backend database and executed there.

That's the plan.

> 2) Connect to a MariaDB server using a native-scheme protocol client,
> and ask the MariaDB server to use CONNECT to proxy individual tables
> from the JDBC backend.
>
> ...is that right?
>
> In which case, I'm afraid to say, option (2) offers only rather poor
> access to JDBC databases.

That was a suggested, but discarded option, for reasons like you
enumerate in your following snipped comments.

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.

- Harold