Small edits to dbt tests based on feedback

This PR makes a number of small edits to our dbt tests as described in https://github.com/ccao-data/data-architecture/issues/371. See that issue for a full list of the changes.

There are a few changes from the list that I haven't made yet. I'm curious to get your take:

Update iasworld_oby_card_between_1_and_100 to exclude condos, and see if the results are accurate (i.e., make sure that there are no non-condo buildings where card >100 makes sense)
- These query results make me doubt that the failures are related to condos with large numbers of cards:

select oby.class, failures.*
from z_dev_jecochr_test_failure.iasworld_oby_card_between_1_and_100 failures
join iasworld.oby oby
    on oby.parid = failures.parid
    and oby.taxyr = failures.taxyr
    and oby.cur = 'Y'
    and oby.deactivat is null

Move iasworld_dweldat_card_proration_rate_not_null from dweldat.user24 to dweldat.external_propct and adjust its filters so that it only checks for prorated PINs (i.e. PINs with tiebacks)
- I tried a few different ways of joining dweldat to pardat and filtering for only PINs with tiebacks but none of them seemed to produce a result set that made sense to me. I'm wondering if maybe my interpretation of the suggestion is incorrect; it seems a bit weird to be comparing a proration rate set on the dwelling level against a tieback set at the parcel level, which is making me think that perhaps I'm getting two different types of proration mixed up. Let me know if I'm misunderstanding the suggestion, and if not (meaning that the join is indeed just very complex) I can split this out into a separate issue and kick it over to you.
Adjust iasworld_dweldat_char_ncu_between_0_and_5 to filter for 212s
- These query results are making me think that this test is actually finding useful things:

select * from z_dev_jecochr_test_failure.iasworld_dweldat_char_ncu_between_0_and_5

Thanks for the detailed review @dfsnow! I left clarifying questions on two of your comments and made fixes based on the rest. This should be ready for another look.

Unrelated to your review, I also decided to split out iasworld_dweldat_sfla_matches_class_definition into individual tests based on class in f124704, which I think makes for a clearer workbook reading experience. Now the description states exactly what the square footage should be for the record's class. You can run just this new set of tests if you'd like to see what the output looks like:

dbt test --store-failures --select \
    iasworld_dweldat_sfla_between_1_and_999_for_class_202
    iasworld_dweldat_sfla_between_1000_and_1800_for_class_203
    iasworld_dweldat_sfla_between_1801_and_25000_for_class_204
    iasworld_dweldat_sfla_between_1_and_2200_for_class_205
    iasworld_dweldat_sfla_between_2201_and_4999_for_class_206
    iasworld_dweldat_sfla_between_1_and_2000_for_class_207
    iasworld_dweldat_sfla_between_3801_and_4999_for_class_208
    iasworld_dweldat_sfla_between_5000_and_50000_for_class_209
    iasworld_dweldat_sfla_between_1_and_10000_for_class_210
    iasworld_dweldat_sfla_between_1_and_40000_for_class_211
    iasworld_dweldat_sfla_between_1_and_20000_for_class_212
    iasworld_dweldat_sfla_between_1_and_10000_for_class_234
    iasworld_dweldat_sfla_between_2001_and_3800_for_class_278
    iasworld_dweldat_sfla_between_1_and_20000_for_class_295

ccao-data / data-architecture

Small edits to dbt tests based on feedback #377