We got a new batch of feedback based on the use of our dbt tests in Rogers. Once these edits are done, we should run an export for West and send it over.
All categories
[x] Update test_qc_year_start to 2024 so that we don't report on any 2023 data
[x] Make sure that class is selected via additional_select_columns on as many tests as possible, since it's useful for workbook readers
[x] Move class from the red "Problematic fields" section of the workbook over to the blue "Unique identifier fields" section
[x] Some issues are addressed at the beginning of the year; data may be out of date
Incorrect values
[x] iasworld_dweldat_fixhalf_between_0_and_6: half baths are pretty heterogenous, so either adjust this test to just make sure fixhalf is greater than 0 or design a heuristic such that it's a multiple of the number of units
[x] Same for iasworld_dweldat_rmbed_sf_between_1_and_15
[x] Same for iasworld_dweldat_sfla_sf_between_1_and_10000
[x] iasworld_dweldat_stories_in_accepted_values should also allow 6.00, 7.00, 8.00, and 9.00
[x] Also update the shared_column_char_type_resd docstring to document these values
[ ] 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)
Clarify with Tia/Dan; I think this test is actually correct
Missing values
[ ] Move iasworld_dweldat_card_proration_rate_not_null from dweldat.user24 to dweldat.external_propct and adjust its filters so that it only checks cards that have tieback PINs
Clarify with Tia/Dan; the rules here are complex
[x] Remove iasworld_dweldat_char_frpl_not_null since a null can indicate no fireplaces
[x] Adjust dweldat garage char not-null tests to filter for rows where user33 (garage size) is not 7 (NONE)
[x] iasworld_dweldat_char_gar_att_not_null
[x] iasworld_dweldat_char_gar_area_not_null
[x] iasworld_dweldat_char_gar_cnst_not_null
[x] Add dweldat test to user34 (garage construction) to ensure it is null if user33 (garage size) is 7 (NONE)
[ ] ~Adjust iasworld_dweldat_char_ncu_between_0_and_5 to filter for 212s~ Actually, this test is useful and has caught an impossible value
[x] Remove iasworld_dweldat_char_porch_not_null and adjust iasworld_dweldat_char_porch_accepted_values to reflect the fact that user30 (porch type) can be null if there is no porch
[x] Remove dweldat tests for user4 since it is deprecated
We got a new batch of feedback based on the use of our dbt tests in Rogers. Once these edits are done, we should run an export for West and send it over.
All categories
test_qc_year_startto 2024 so that we don't report on any 2023 dataclassis selected viaadditional_select_columnson as many tests as possible, since it's useful for workbook readersclassfrom the red "Problematic fields" section of the workbook over to the blue "Unique identifier fields" sectionIncorrect values
iasworld_dweldat_fixhalf_between_0_and_6: half baths are pretty heterogenous, so either adjust this test to just make sure fixhalf is greater than 0 or design a heuristic such that it's a multiple of the number of unitsiasworld_dweldat_rmbed_sf_between_1_and_15iasworld_dweldat_sfla_sf_between_1_and_10000iasworld_dweldat_stories_in_accepted_valuesshould also allow 6.00, 7.00, 8.00, and 9.00shared_column_char_type_resddocstring to document these valuesiasworld_oby_card_between_1_and_100to 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)Missing values
iasworld_dweldat_card_proration_rate_not_nullfromdweldat.user24todweldat.external_propctand adjust its filters so that it only checks cards that have tieback PINsiasworld_dweldat_char_frpl_not_nullsince a null can indicate no fireplacesiasworld_dweldat_char_gar_att_not_nulliasworld_dweldat_char_gar_area_not_nulliasworld_dweldat_char_gar_cnst_not_nulliasworld_dweldat_char_ncu_between_0_and_5to filter for 212s~ Actually, this test is useful and has caught an impossible valueiasworld_dweldat_char_porch_not_nulland adjustiasworld_dweldat_char_porch_accepted_valuesto reflect the fact that user30 (porch type) can be null if there is no porchiasworld_dweldat_char_tp_dsgn_not_nulliasworld_dweldat_char_tp_dsgn_in_accepted_valuesiasworld_dweldat_fixhalf_not_nullsince half baths can be null if there are no half baths