feat(tableau): add is_completion_row to ddi_dashboard

[anthonygwalters]

Adds a boolean flag that is TRUE for the single deduplication row per student-assessment: response_type = 'overall' for completed assessments, or NULL response_type for assigned-but-incomplete assessments.

Filtering on is_completion_row = TRUE instead of response_type = 'overall' exposes incomplete assessments to Tableau, enabling a simple SUM(is_complete) / COUNT(is_complete) completion rate that responds naturally to any dimension filter (subgroup, school, grade, etc.) without requiring LOD expressions.

Pull Request

Summary & Motivation

"When merged, this pull request will..."

AI Assistance

If Claude Code authored or co-authored this PR, briefly note what was
AI-assisted vs. human-directed in the summary above

Self-review

Complete only the sections relevant to your changes.

General

If this is a same-day request, please flag that in the #data-team Slack

• Update due date and assignee on the
  TEAMster Asana Project
• Review the Claude Code Review comment posted on this PR. Address valid
  feedback; dismiss false positives with a brief reply explaining why.
  (Claude is advisory — use your judgement, but don't ignore it.)

Dagster (skip if no Dagster changes)

• Run uv run dagster definitions validate for any modified code location
• Run uv run pytest tests/test_dagster_definitions.py for any modified
  code location
• New integrations follow the
  Library + Config pattern
  with the correct asset key format and IO manager

dbt (skip if no dbt changes)

• Include a [model_name].yml properties file for all new models — see
  dbt Conventions
• Include (or update) an
  exposure
  for all models consumed by a dashboard, analysis, or application
• Breaking change? Renaming or removing columns in a contracted model
  (stg_, rpt_, mart) will break downstream exposures. Coordinate with
  affected teams before merging and document your rollback plan in the
  summary above.
• If adding a new external source, run stage_external_sources before
  building — see
  SQL conventions

Docs (skip if no schedule, sensor, or integration changes)

• If adding or changing a schedule or sensor, regenerate the automations
  catalog: uv run scripts/gen-automations-doc.py
• If adding a new integration to a code location, update that location's
  CLAUDE.md by running /claude-md-management:revise-claude-md

CI checks

• Trunk — passes. If it fails, run trunk check and trunk fmt
  locally, fix any issues, and push again.
• dbt Cloud — passes. If it fails: click Details on the check →
  expand Invoke dbt build ... → Debug Logs. Or tag Claude on
  the PR with the error details.
• Dagster Cloud — passes or not triggered (only runs when relevant paths
  change and the PR is not a draft). If it fails: check the Actions tab
  for the workflow run logs. Re-run failed jobs if the failure looks
  transient; otherwise check the Dagster Cloud branch deployment for
  error details. Or tag Claude on the PR for help.

Troubleshooting

• SqlFluff Rules Reference
• Dagster "kinds" Reference
• Automated checks
• dbt Conventions

[github-actions]

Claude finished @anthonygwalters's task in 1m 44s —— View job

---

Code Review

• Gather context
• Read changed files
• Review dbt SQL and YAML changes
• Post review feedback

---

Summary

This is a clean, additive change. The logic is sound and the documentation is well-written. A few observations below.

---

SQL — rpt_tableau__ddi_dashboard.sql

Logic — correct ✓

if(
    co.response_type = 'overall' or co.response_type is null, true, false
) as is_completion_row,

The two-branch condition correctly maps the stated semantics:

• 'overall' → completed assessment deduplication row
• NULL → assigned-but-incomplete assessment (no response rows exist, so the join produces NULL)

UNION ALL consistency — correct ✓

is_completion_row appears in all three branches of the UNION ALL:

• ES/MS block: computed (line 273)
• HS block: computed (line 381)
• Walkthrough block: null (line 512)

Using null (not false) for walkthroughs is consistent with the existing convention for boolean fields that don't apply to walkthroughs (is_complete, is_qbl, etc.), and it correctly excludes walkthrough rows from WHERE is_completion_row = TRUE filtering.

Minor: is_complete vs is_completion_row type inconsistency (pre-existing)

is_complete returns INT64 (if(..., 1, 0)) while is_completion_row returns BOOLEAN (if(..., true, false)). This is a pre-existing pattern — not introduced here — but worth noting for Tableau consumers who might expect consistency.

---

YAML — rpt_tableau__ddi_dashboard.yml

Column entry — correct ✓

- name: is_completion_row
  data_type: boolean
  description: >-
    True for the single deduplication row per student-assessment: ...

data_type: boolean matches the IF() return type in BigQuery. The description is accurate and clearly explains the filter pattern for Tableau consumers.

Convention gap: missing description: on model and existing columns

Per src/dbt/CLAUDE.md: "All new or modified models require description: on the model and every column." This model is being modified, so a model-level description: and descriptions for all existing columns are technically required. This is a pre-existing gap across ~70 columns — not introduced by this PR — so I'm flagging it as an observation rather than a blocker. If there's a convention to handle this incrementally, that would apply here.

---

Exposure — no changes needed ✓

The ddi_suite exposure in tableau.yml already references ref("rpt_tableau__ddi_dashboard") and doesn't enumerate specific columns, so no update is needed.

---

Assessment

Ready to merge. The implementation is correct, consistent, and appropriately documented. The convention gap on column descriptions is pre-existing and doesn't affect correctness or the Tableau use case.

[github-actions]

Claude finished @anthonygwalters's task in 3m 0s —— View job

---

Code Review

• Gather context
• Read changed files
• Review dbt SQL and YAML changes
• Post review feedback

---

Summary

Clean, purposeful change. The logic is correct and the description is excellent. One fix requested before merge.

---

SQL — rpt_tableau__ddi_dashboard.sql

Logic — correct ✓

if(
    co.response_type = 'overall' or co.response_type is null, true, false
) as is_completion_row,

Both conditions are semantically correct: 'overall' is the right deduplication row for completed assessments; NULL response_type correctly identifies assigned-but-incomplete assessments (no response rows exist, so the join produces NULL).

UNION ALL consistency — correct ✓

is_completion_row appears consistently in all three branches and is positioned immediately after is_qbl in both the ES/MS (line 273) and HS (line 381) blocks. The walkthrough block correctly uses null as is_completion_row (line 512), consistent with the existing convention for flags not applicable to walkthroughs (is_complete, is_qbl, etc.).

---

YAML — rpt_tableau__ddi_dashboard.yml — fix requested

Column description — correct ✓

data_type: boolean matches the IF() return type. The description is accurate and written from the consumer's perspective — exactly what it should be.

YAML column order does not match SQL SELECT order

is_completion_row is appended at the end of the YAML column list (after is_low_25_fl), but in the SQL SELECT it appears between is_qbl and is_passed_iready_2plus_reading_int. BigQuery contract enforcement matches columns by name, not position, so this won't cause a runtime failure — but the YAML should reflect SQL ordering for maintainability.

Please move the is_completion_row entry in the YAML to appear between is_qbl and is_passed_iready_2plus_reading_int:

      - name: is_qbl
        data_type: boolean
      - name: is_completion_row          # ← move here
        data_type: boolean
        description: >-
          True for the single deduplication row per student-assessment: the
          response_type = 'overall' row for completed assessments, or the null
          response_type row for assigned-but-incomplete assessments. Filter on
          this field instead of response_type = 'overall' when computing
          completion rates so that incomplete assessments are included in the
          denominator.
      - name: is_passed_iready_2plus_reading_int
        data_type: int64

Note: There is a broader pre-existing YAML/SQL ordering mismatch in this file (is_foundations, is_self_contained, year_in_network, is_sipps, is_low_25_fl are also out of order relative to the SQL) — that's out of scope for this PR, but this PR shouldn't deepen it by appending at the end.

---

Exposure — no changes needed ✓

The ddi_suite exposure already references ref("rpt_tableau__ddi_dashboard") without enumerating specific columns.

---

Assessment

One fix before merge — move is_completion_row in the YAML to match its SQL SELECT position. Everything else is correct.

[github-actions]

Claude finished @cbini's task in 2m 0s —— View job

---

Code Review

• Gather context
• Read changed files
• Review dbt SQL and YAML changes
• Post review feedback

---

Summary

This is the fix requested in the previous review — the change itself is correct. One outstanding item remains before merge.

---

SQL — correct ✓

is_completion_row logic is sound and UNION ALL placement is consistent across all three branches (ES/MS line 273, HS line 381, walkthrough null line 512).

---

YAML — fix applied locally, push blocked

The previous review correctly flagged that is_completion_row was appended at the end of the YAML column list, but in the SQL SELECT it appears between is_qbl and is_passed_iready_2plus_reading_int. I applied the fix locally (commit 685147c) but the push was rejected (GitHub Actions bot lacks write permission to this repo).

Please apply this fix manually — move the is_completion_row entry in the YAML from the end to between is_qbl and is_passed_iready_2plus_reading_int:

      - name: is_qbl
        data_type: boolean
      - name: is_completion_row          # ← move here
        data_type: boolean
        description: >-
          True for the single deduplication row per student-assessment: the
          response_type = 'overall' row for completed assessments, or the null
          response_type row for assigned-but-incomplete assessments. Filter on
          this field instead of response_type = 'overall' when computing
          completion rates so that incomplete assessments are included in the
          denominator.
      - name: is_passed_iready_2plus_reading_int
        data_type: int64

And remove it from the end of the file (after is_low_25_fl).

---

Assessment

One fix before merge — YAML column ordering. Everything else is correct and the implementation logic is solid.