Best schema for a big inventory doc

I’m trying to create a inventory doc in Coda that deals with its limitation.
Coda can’t handle big tables well. It’s getting better each day but it’s not there yet.

This template above shows the traditional implementation with a table called “Order Item Qty”. The problem with this approach is that for a single order you have to create n rows for n items as you can see in the picture below. It adds up really fast.

My first attempt the improve this schema was to get rid of this “Order Item Qty” table and create a v2 of Purchases Orders using arrays to save data. I can always access and save data using Nth() and ListCombine() function.
My Purchases Orders Table would look like this:

With this v2 you can create only one row for each order but it’s very difficult to do backup because of the lookup columns that loses their references.
My idea now is to get rid of the lookup columns and implement the v3 bellow:

Am I on the right track?
Is there a better way to do big inventory in Coda?
What are the approaches people are using to save large amount of data?
ps. I already got rid of button columns and conditional format.

I have a large inventory doc and I use your first method. I have found that there is info I need at the line item level that a specific table is the best solution. I then aggregate to invoice/order as you are doing it.

Since the Coda storage limits are still rather opaque its hard to bend the schemas … perhaps keep your Order Items table as simple as possible as Coda seems to store less complicated tables compactly.

Thank you for your help, @Johg_Ananda.
I think that wouldn’t work for me. My tables would be bloated with more than 10.000 rows easily.
I’d be really glad if more people commented on this.