add same rule for historic obj assess (#103)

models/core_warehouse/fct_student_objective_assessment.sql

@@ -15,6 +15,9 @@ with student_obj_assessments_long_results as (
 student_obj_assessments as (
     select * from {{ ref('stg_ef3__student_objective_assessments') }}
 ),
+dim_student as (
+    select * from {{ ref('dim_student') }}
+),
 object_agg_other_results as (
     select
         k_student_objective_assessment,
@@ -29,9 +32,10 @@ student_obj_assessments_wide as (
         student_obj_assessments.k_objective_assessment,
         student_obj_assessments.k_student_assessment,
         student_obj_assessments.k_assessment,
-        student_obj_assessments.k_student,
+        dim_student.k_student,
+        student_obj_assessments.k_student_xyear,
         student_obj_assessments.tenant_code,
-        school_year,
+        student_obj_assessments.school_year,
         administration_date,
         administration_end_date,
         event_description,
@@ -59,7 +63,15 @@ student_obj_assessments_wide as (
         and student_obj_assessments_long_results.normalized_score_name != 'other'
     left join object_agg_other_results
         on student_obj_assessments.k_student_objective_assessment = object_agg_other_results.k_student_objective_assessment
-    {{ dbt_utils.group_by(n=18) }}
+    -- left join to allow 'historic' records (assess records with no corresponding stu demographics)
+    left join dim_student
+        on student_obj_assessments.k_student = dim_student.k_student
+    -- FILTER to students who EVER have a record in dim_student
+    where student_obj_assessments.k_student_xyear in (
+        select distinct k_student_xyear
+        from dim_student
+    )
+    {{ dbt_utils.group_by(n=19) }}
 )
 select *
 from student_obj_assessments_wide

models/core_warehouse/fct_student_objective_assessment.yml

@@ -3,9 +3,27 @@ version: 2
 models: 
   - name: fct_student_objective_assessment
     description: >
-      Student objective assessment records. The wide results columns are
-      determined based on the `normalized_score_name` column
-      from the seed table `xwalk_objective_assessment_scores`.
+      ##### Overview:
+        This fact table defines student objective assessment records.
+
+      ##### Primary Key:
+        `k_student_objective_assessment` -- There is one record per student-objective assessment event
+
+      ##### Important Business Rules:
+        The wide score results columns are determined based on the `normalized_score_name`
+        column from the seed table `xwalk_assessment_scores`.
+
+        `k_student` is an annualized student identifier, so it will be null unless there is a student record for the relevant school year.
+        We INCLUDE records with null `k_student` to enable analysis of current students' full history of assessments. Therefore,
+        users CAN use `fct_student_objective_assessment` to answer "What is the longitudinal trend in test scores for my current students?"
+        
+        However, `k_student_xyear` is a x-year student identifier, and we EXCLUDE records where `k_student_xyear` is null,
+        because we have no way to identify the students. Therefore, users CAN NOT use `fct_student_objective_assessment` to answer 
+        "What is the longitudinal trend in test scores for students who were english learners at the time of the test?",
+        UNLESS every year of assessment data has a corresponding & complete set of identifying data in `dim_student`.
+        Often, this means you should only run longitudinal analysis on years for which you have populated an Ed-Fi ODS, 
+        or a supplementary data source.
+
     config:
       tags: ['assessment']
       enabled: "{{ var('src:domain:assessment:enabled', True) }}"