Particular highight of both rows and columns

I am analysing a large data worksheet (Sheet1) which consists of 21 columns (A-U) and over 35000 rows.
Data are sorted in groups of contiguous rows.
All data groups are contiguous, ie no empty rows separating them.
ColP reflects in its cells the number of rows of each group. So if a group is made up of three rows, ColP shows in each cell in its column, of that group, number 3 as demonstrated in Sheet1 of the attached example.

I am in need to:

1. highlight each group of rows alternately in two different colours as per Sheet2 of the example, and
2. highlight, in a third colour, ColP cells for the data group which are not in the same number as shown in those cells. As per Sheet2 of the example, cells P4-P12 are highlighed in blue because the group is of 9 rows not 11. The same applied on cells P24-P39 because the group is of 16 rows not 17 as shown in ColP.

Can I get some expert assistance with this need, please?

Many thanks in advance.

T.

Excellent work Karedog. When I ran the code on my sent example, it did what is required, but actually highlighted cells P13-P16 which are not needed to be highlighted as there's nothing wrong with them.
Could you please run it on my example to see what I mean.
I'm always more than obliged for your precious assistance.
Thank you very much indeed.
T.

Your are 100% right. But actually, they are two groups of 2 rows. They are highlighted in Sheet2 in two colours.
Now you're making me think how to differentiate between the correct groups and the mistaken ones?
Well, in this case I say I will run your code on the real data sheet, and as I will filter by ColP highlighted cells, I will ignore what I consider correct groups.
In other words, this problem is solved and added to your "ammunition" of genius code.
Thank you very very much Karedog for your constant precious help. You left me speechless.

All the best ..
T.

Yes .. definitely yes .. you did it again Karedog .. 100% right .. nothing wrong at all ..

Always .. Thank you my dear ..

T.