VK9 | Direct3D 9 compatibility layer using Vulkan | Game Engine library

I have the below 2 tables:

t1

VK AY AN R VK1 A1, A2 1 VK2 2 VK3 A1, A2, A3, A4 3 VK4 A2 4 VK5 5 VK6 6 VK7 A3 7 VK8 8 VK9 9 VK10 10 VK11 11 VK12 12 VK13 A3 13 VK14 14 VK15 A3 15 VK16 16 VK17 17 VK18 18 VK19 19 VK20 20

t2

A C A1 4 A2 10 A3 2 A4 10

I would like to create a plsql cursor to add a column AF with the logic below:

1. The objective is to assign the different As in t2 to the rows in t1. It's column R which determines the order in which, if possible, they should be assigned.
2. If column AY is populated, then that 'A' should be assigned to AF if C in t2 is still >0 (after the previous assignations).
3. If column AN is populated, then none of the 'A's listed there can be assigned to that row.
4. The assignation of 'A's must be done proportionally to their C in t2.

The below would be one example of desired output (AF could be assigned in any other order though).

VK AY AN R AF EXPLANATION VK1 A1, A2 1 A4 A3 and A4 available: A4 chosen. VK2 2 A2 All available: A2 chosen. VK3 A1, A2, A3, A4 3 null None are available: cannot pick any. VK4 A2 4 A4 All but A2 available: A4 chosen VK5 5 A2 All available: A2 chosen. VK6 6 A1 All available: A1 chosen. VK7 A3 7 A3 A3 must be assigned: A3 chosen VK8 8 A2 All available: A2 chosen. VK9 9 A4 All available: A4 chosen. VK10 10 A4 All available: A4 chosen. VK11 11 A2 All available: A2 chosen. VK12 12 A1 All available: A1 chosen. VK13 A3 13 A3 A3 must be assigned: A3 chosen VK14 14 A2 All available: A2 chosen. VK15 A3 15 null A3 must be assigned but A3 already exhausted: cannot be assigned VK16 16 A4 All available: A4 chosen. VK17 17 A4 All available: A4 chosen. VK18 18 A2 All available: A2 chosen. VK19 19 A2 All available: A2 chosen. VK20 20 A1 All available: A1 chosen.

Please find below the sample query for the tables:

...

I'd like to use silhouette_score to estimate the optimum number of clusters. I'm using an official example from sklearn but it gives me this error: TypeError: silhouette_score() takes 1 positional argument but 2 were given.

My data (X) is a pandas dataframe with 20 features (all non-null float64) and index as unique ID strings (can this be a problem?).

...