Merged PR 2743: Fixes deal-based issues on the billing country dimension

# Description

Before deploying KPIs by Billing Country, we spotted some issues that were basically increases on the volumes of any metric on the by billing country dimension that was based on Deal. This means, `int_core__mtd_deal_metrics` and `int_xero__mtd_invoicing_metrics`.

This PR changes the following:
* Now the 2 abovementioned models depend on the `int_core__deal` model, instead of `int_core__user_host` (thus removing duplicated stuff)
* Now all models use the main billing country at deal level, instead of doing it so at host level. The reason is that some small amount of hosts that share the same deal can have a different billing country. To avoid weird stuff, everything points to this simplification - that in general, it's not a massive change in the output.
* In order to do so easily, the 3 main billing country per deal fields have been propagated to `int_core__user_host`

To exemplify the solution, find here a snapshot of the differences in behavior:

```
select
    dimension,
    sum(deals_booked_in_month) as deals_booked_1,
    sum(deals_booked_in_6_months) as deals_booked_6,
    sum(deals_booked_in_12_months) as deals_booked_12,
    sum(total_revenue_in_gbp) as total_revenue,
    sum(xero_operator_net_fees_in_gbp) as operator_revenue,
    sum(xero_booking_net_fees_in_gbp) as booking_fees,
    sum(xero_listing_net_fees_in_gbp) as listing_fees,
    sum(xero_verification_net_fees_in_gbp) as verification_fees,
    sum(total_guest_revenue_in_gbp) as guest_revenue,
    sum(xero_waiver_paid_back_to_host_in_gbp) as waiver_paid_back_to_hosts,
    sum(waiver_net_fees_in_gbp) as waiver_net_fees
from intermediate.int_mtd_vs_previous_year_metrics
where date in ('2024-01-31')
group by 1
order by 1
```
Production:
![image.png](https://guardhog.visualstudio.com/4148d95f-4b6d-4205-bcff-e9c8e0d2ca65/_apis/git/repositories/54ac356f-aad7-46d2-b62c-e8c5b3bb8ebf/pullRequests/2743/attachments/image.png)

vs.
Local:
![image (2).png](https://guardhog.visualstudio.com/4148d95f-4b6d-4205-bcff-e9c8e0d2ca65/_apis/git/repositories/54ac356f-aad7-46d2-b62c-e8c5b3bb8ebf/pullRequests/2743/attachments/image%20%282%29.png)

Keep in mind that still Global dimension can be greater than any other dimension aggregated since not all users have a deal. Mismatches between the other 2 dimensions might be linked to the dump.

Commits are meaningful and help navigate in the changes.

# Checklist

- [X] The edited models and dependants run properly with production data.
- [X] The edited models are sufficiently documented.
- [X] The edited models contain PK tests, and I've ran and passed them.
- [X] I have checked for DRY opportunities with other models and docs.
- [X] I've picked the right materialization for the affected models.

# Other

- [ ] Check if a full-refresh is required after this PR is merged.

Related work items: #20823
This commit is contained in:
Oriol Roqué Paniagua 2024-09-05 09:53:16 +00:00
parent bc77e5df08
commit 435db55c1e
10 changed files with 49 additions and 26 deletions

View file

@ -31,7 +31,7 @@ with
{% elif dimension.dimension == "'by_billing_country'" %}
inner join int_core__user_host h
on d.date >= h.created_date_utc
and h.billing_country_iso_3 is not null
and h.main_billing_country_iso_3_per_deal is not null
{% endif %}
{% if not loop.last %}
union all