Validated queries

Think of validated queries as training data or examples that Qluent uses with AI to better answer your questions.

Use case

In cases where Qluent struggles to answer a common question it can be fixed by adding a validated example, providing that you know the SQL code. Qluent is smart enough to adapt the given SQL code to different questions therefore it's not necessary to provide all variations. e.g. Query Can you give me the week over week growth for sessions for weeks 41-44? and code

WITH weekly_sessions AS (
    SELECT 
        DATE_PART('week', date) AS week,
        SUM(sessions) AS total_sessions
    FROM google_analytics_daily
    WHERE DATE_PART('year', date) = 2024
    AND DATE_PART('week', date) BETWEEN 40 AND 44 
    GROUP BY DATE_PART('week', date)
    ORDER BY DATE_PART('week', date)
)
SELECT 
    week,
    LAG(total_sessions) OVER (ORDER BY week) AS previous_week_sessions,
    ((total_sessions::DECIMAL - LAG(total_sessions) OVER (ORDER BY week)::DECIMAL) 
        / NULLIF(LAG(total_sessions) OVER (ORDER BY week)::DECIMAL, 0)) * 100 AS week_over_week_growth
FROM weekly_sessions
WHERE week BETWEEN 41 AND 44