Skip to content

Create data pipeline: gold_user_engagement #6

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
zingleai wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Create data pipeline: gold_user_engagement #6

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 10 additions & 0 deletions dbt/models/marts/gold_user_engagement.sql
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,10 @@
-- Model: gold_user_engagement
-- User engagement metrics by channel for Marketing dashboard
SELECT
channel,
COUNT(DISTINCT user_id) AS active_users,
SUM(sessions) AS total_sessions,
AVG(session_duration_minutes) AS avg_session_duration
FROM {{ ref('stg_user_activity') }}
WHERE activity_date >= DATEADD(day, -7, CURRENT_DATE)
Copy link
Copy Markdown

@coderabbitai coderabbitai bot Feb 23, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

⚠️ Potential issue | 🟠 Major

Off-by-one: >= DATEADD(day, -7, CURRENT_DATE) spans 8 days, not 7.

With today as the anchor, DATEADD(day, -7, CURRENT_DATE) resolves to 7 days ago. Using >= therefore includes that anchor day plus 7 subsequent days (including today) — a total of 8 days. Additionally, aggregating today's still-accumulating data skews the daily metrics for the Marketing dashboard.

For 7 complete days (exclusive of today):

WHERE activity_date >= DATEADD(day, -7, CURRENT_DATE)
  AND activity_date < CURRENT_DATE

Or, for a rolling 7-day window ending today inclusive but exactly 7 days:

WHERE activity_date >= DATEADD(day, -6, CURRENT_DATE)

Choose whichever matches the business requirement.

🤖 Prompt for AI Agents 
Verify each finding against the current code and only fix it if needed.

In `@dbt/models/marts/gold_user_engagement.sql` at line 9, The WHERE clause using
activity_date >= DATEADD(day, -7, CURRENT_DATE) is off-by-one and yields 8 days
(includes today); update the filter for the intended 7-day window by either (a)
adding an upper bound activity_date < CURRENT_DATE to get 7 complete days
excluding today, or (b) shifting the lower bound to DATEADD(day, -6,
CURRENT_DATE) to include today as the 7th day—modify the WHERE clause
referencing activity_date, DATEADD and CURRENT_DATE accordingly to match the
business requirement.

⚠️ Potential issue | 🟠 Major

🧩 Analysis chain

🌐 Web query:

dbt dateadd macro documentation

💡 Result:

dateadd is now documented as a dbt Core cross-database macro (namespace: dbt), not in dbt_utils (those cross-db macros were removed from dbt_utils starting in v1.0). [1] (docs.getdbt.com)

dbt dateadd macro (Core)

Signature

{{ dbt.dateadd(datepart="day", interval=1, from_date_or_timestamp="..." ) }}

Args

  • datepart: a date/time part (database-specific)
  • interval: integer (positive or negative)
  • from_date_or_timestamp: a date/time expression [1] (docs.getdbt.com)

Examples

{{ dbt.dateadd(datepart="day", interval=1, from_date_or_timestamp="'2016-03-09'") }}
{{ dbt.dateadd(datepart="month", interval=-2, from_date_or_timestamp="'2016-03-09'") }}
``` [1] ([docs.getdbt.com](https://docs.getdbt.com/reference/dbt-jinja-functions/cross-database-macros))

### Sources
- [1] dbt Developer Hub — “Cross-database macros” (includes `dateadd`) ([docs.getdbt.com](https://docs.getdbt.com/reference/dbt-jinja-functions/cross-database-macros))  
- [2] dbt-utils README note about cross-database macros moving to dbt Core ([github.com](https://github.com/dbt-labs/dbt-utils?utm_source=openai))

Citations:

- 1: https://docs.getdbt.com/reference/dbt-jinja-functions/cross-database-macros
- 2: https://docs.getdbt.com/reference/dbt-jinja-functions/cross-database-macros
- 3: https://docs.getdbt.com/reference/dbt-jinja-functions/cross-database-macros
- 4: https://docs.getdbt.com/reference/dbt-jinja-functions/cross-database-macros
- 5: https://github.com/dbt-labs/dbt-utils?utm_source=openai

---

🌐 Web query:

`DATEADD function Snowflake Redshift PostgreSQL BigQuery compatibility`

💡 Result:

### `DATEADD` compatibility overview (Snowflake vs Redshift vs PostgreSQL vs BigQuery)

| Engine | Has `DATEADD`? | Add *N* units to a date/time |
|---|---:|---|
| **Snowflake** | Yes | `DATEADD(part, value, expr)` (aliases: `TIMEADD`, `TIMESTAMPADD`) [1] |
| **Amazon Redshift** | Yes | `DATEADD(datepart, interval, expr)` [2] |
| **PostgreSQL** | **No** (built-in) | Use arithmetic: `timestamp + interval '1 day'`, `date + integer` (days), etc. [5] |
| **BigQuery (GoogleSQL)** | **No** (`DATEADD` name) | Use typed functions: `DATE_ADD(date, INTERVAL n part)`, `DATETIME_ADD(...)`, `TIMESTAMP_ADD(...)`, `TIME_ADD(...)` [3][4] |

---

## Syntax mappings

### Snowflake → Redshift (nearly 1:1)
Both accept the same 3-argument pattern: `(part, number, expression)` [1][2].

```sql
-- Snowflake / Redshift
DATEADD(day, 7, some_ts)

Snowflake / Redshift → BigQuery

BigQuery uses *_ADD(expr, INTERVAL n PART) (note the INTERVAL clause and PART position) [3][4]. Google’s Snowflake migration guide summarizes this mapping as:

  • DATEADD(part, value, dateOrTime) → DATE_ADD(date, INTERVAL value part) [6]
-- BigQuery
DATE_ADD(some_date, INTERVAL 7 DAY)
TIMESTAMP_ADD(some_ts, INTERVAL 7 DAY)

Snowflake / Redshift → PostgreSQL

PostgreSQL typically uses + interval (and for date, + integer adds days) [5].

-- PostgreSQL
some_ts + interval '7 day'
some_date + 7

Key behavioral gotchas (portability)

  • Return type differences when input is DATE and you add sub-day units

    • Snowflake: adding hour/minute/second to a DATE returns a timestamp (TIMESTAMP_NTZ) starting at midnight [1].
    • Redshift: DATEADD returns TIMESTAMP for date inputs (date becomes timestamp at 00:00:00) [2].

  • Month/quarter/year end-of-month behavior

    • BigQuery: for MONTH/QUARTER/YEAR, if the target month is shorter, the result becomes the last day of that month [3].
    • Snowflake: warns that year/quarter/month can shift the day-of-month when the result month has fewer days [1].
    • PostgreSQL interval math explicitly uses “last day of month” when needed while applying the months field of an interval [5].

  • BigQuery TIMESTAMP_ADD supports only up to DAY

    • TIMESTAMP_ADD(..., INTERVAL n DAY/HOUR/MINUTE/...) is supported, but not MONTH/QUARTER/YEAR for TIMESTAMP_ADD [4]. (Use DATETIME_ADD, or convert to DATETIME/DATE where appropriate.)

Sources: Snowflake DATEADD [1], Redshift DATEADD [2], BigQuery DATE_ADD [3], BigQuery TIMESTAMP_ADD [4], PostgreSQL date/time operators & interval rules [5], BigQuery Snowflake translation guide [6].

Use the {{ dbt.dateadd() }} macro instead of DATEADD for cross-warehouse compatibility.

DATEADD(datepart, interval, date) is native to Snowflake and Redshift but PostgreSQL does not have a built-in DATEADD function, and BigQuery uses DATE_ADD/DATETIME_ADD/TIMESTAMP_ADD. dbt provides a cross-warehouse {{ dbt.dateadd() }} macro that compiles to the correct dialect for each adapter.

🔧 Proposed fix using dbt's built-in macro 
-WHERE activity_date >= DATEADD(day, -7, CURRENT_DATE)
+WHERE activity_date >= {{ dbt.dateadd(datepart="day", interval=-7, from_date_or_timestamp=current_date) }}
📝 Committable suggestion

‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.

Suggested change
WHERE activity_date >= DATEADD(day, -7, CURRENT_DATE)
WHERE activity_date >= {{ dbt.dateadd(datepart="day", interval=-7, from_date_or_timestamp="CURRENT_DATE") }}
🤖 Prompt for AI Agents 
Verify each finding against the current code and only fix it if needed.

In `@dbt/models/marts/gold_user_engagement.sql` at line 9, Replace the
warehouse-specific DATEADD usage in the WHERE clause with dbt's cross-warehouse
macro: locate the WHERE filtering on activity_date that currently uses
DATEADD(day, -7, CURRENT_DATE) and call the dbt.dateadd macro instead (keep the
same datepart "day", interval -7, and CURRENT_DATE as the date argument) so the
compiled SQL will work across adapters; update the expression that references
activity_date and CURRENT_DATE accordingly.

GROUP BY 1
Comment on lines +1 to +10
Copy link
Copy Markdown

@coderabbitai coderabbitai bot Feb 23, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

⚠️ Potential issue | 🟠 Major

Missing {{ config() }} block — model will not materialize as a table.

The PR metadata specifies materialized: table, but there is no {{ config() }} block in the file. Without it, dbt uses the project-level default (typically view), silently mismatching the intended materialisation.

🔧 Proposed fix 
+{{ config(materialized='table') }}
+
 -- Model: gold_user_engagement  
 -- User engagement metrics by channel for Marketing dashboard
🤖 Prompt for AI Agents 
Verify each finding against the current code and only fix it if needed.

In `@dbt/models/marts/gold_user_engagement.sql` around lines 1 - 10, The model
gold_user_engagement is missing a dbt config block so it will default to the
project-level materialization (likely view) instead of the intended table; add a
top-of-file {{ config(materialized='table') }} block (before the SELECT) in the
gold_user_engagement model to explicitly set materialized: table, leaving the
rest of the query (including ref('stg_user_activity')) unchanged.