What get method?

Of the Query interface (I'm looking at the github readme)

Aha. Yes, it does execute it for every row, but you could just accumulate the data into a construct that returns both the customer and the order, then easily tie them together with a single line of code later.

For example, the query could return a

Pair<Customer, Order>

That's exactly what I'm trying to avoid haha. Looking at the KDBC example code I'd implement what I want like this: In the init block I'd have a group(c.id()) function that tells KDBC to group the results on the customer id. In the get() method I'd then receive each grouped list of rows which I could then easily transform into a Customer with a list of Orders.

Or, you could have a list of orders inside Customer, and use that.

Aha, so just a shortcut to this mechanism. That wouldn't be too hard.

Yes, it's just that with Jooq or JDBC you have to iterate the ResultSet yourself to group the rows which is boilerplate so having KDBC do that for me would be great

I can make an issue for it if needed

Cool! I would like to add this feature for sure!

I think this would be the manual approach:

@User uploaded a file: Untitled

Did test it, just wrote it out.

I assume that customer has an

orders

list, and I used a cartesian product to get the customer and all the orders

This is too much cognitive effort for sure, so I really like your idea.

Not more time tonight though.

There's no urgency. Its just something that has been on the back of my mind for a while since I encountered the problem in a project. I'm not in a situation where I immediately need it.

@teedee Thanks for creating the issue 🙂

@teedee I looked at it a little now. My first attempt was to create allow a query for

Pair<Customer, Order>

into

Map<Customer, List<Order>>

with a single command.

@User uploaded a file: Query for Customers and Orders (cartesian product)

To run this query and group all orders by customer you'd do:

@User uploaded a file: Untitled

I added the type on

customers

to make it clearer.

This is the generated SQL:

@User uploaded a file: Untitled

Note that the aliases are optional, KDBC will make sure there are no ambiguities even if you don't add aliases manually, I just did it to keep the output at a minimum. *It would prefix each column with the table name by default)

The aliases will automatically be mapped to the correct places in the table objects of course.

That's a good start but ideally you'd have Query<Customer> where calling list() would be sufficient to get the list of Customer objects with the list of orders as a field. If KDBC is built with the assumption of get() handling only a single row that might be a little more complex however.

@teedee This was only the first step, as all I did was add an extension function on

Iterable<Pair<K, V>>

. Maybe we could introduce an intermediary step that can be run before

map

, for example

aggregate

?