십분위수 기반 파생 데이터 세트 사용 사례

CREATE TABLE AS airline_loyality_decile
{  WITH summed_miles_1 AS (
        SELECT _profilefoundationreportingstg.membershipNumber AS membershipNumber,
            _profilefoundationreportingstg.loyaltyStatus AS loyaltyStatus,
            SUM(_profilefoundationreportingstg.mileage) AS totalMiles
        FROM airline_loyalty_data
        WHERE _profilefoundationreportingstg.transactionDate < (MAKE_DATE(YEAR(CURRENT_DATE), MONTH(CURRENT_DATE), 1) - MAKE_YM_INTERVAL(0, 0))
    GROUP BY 1,2
    ),
    summed_miles_3 AS (
        SELECT _profilefoundationreportingstg.membershipNumber AS membershipNumber,
            _profilefoundationreportingstg.loyaltyStatus AS loyaltyStatus,
            SUM(_profilefoundationreportingstg.mileage) AS totalMiles
        FROM airline_loyalty_data
        WHERE _profilefoundationreportingstg.transactionDate < (MAKE_DATE(YEAR(CURRENT_DATE), MONTH(CURRENT_DATE), 1) - MAKE_YM_INTERVAL(0, 1))
    GROUP BY 1,2
    ),
    summed_miles_6 AS (
        SELECT _profilefoundationreportingstg.membershipNumber AS membershipNumber,
            _profilefoundationreportingstg.loyaltyStatus AS loyaltyStatus,
            SUM(_profilefoundationreportingstg.mileage) AS totalMiles
        FROM airline_loyalty_data
        WHERE _profilefoundationreportingstg.transactionDate < (MAKE_DATE(YEAR(CURRENT_DATE), MONTH(CURRENT_DATE), 1) - MAKE_YM_INTERVAL(0, 4))
    GROUP BY 1,2
    ),
    rankings_1 AS (
        SELECT membershipNumber,
            loyaltyStatus,
            totalMiles,
            NTILE(10) OVER (PARTITION BY loyaltyStatus ORDER BY totalMiles DESC) AS decileBucket
        FROM summed_miles_1
    ),
    rankings_3 AS (
        SELECT membershipNumber,
            loyaltyStatus,
            totalMiles,
            NTILE(10) OVER (PARTITION BY loyaltyStatus ORDER BY totalMiles DESC) AS decileBucket
        FROM summed_miles_3
    ),
    rankings_6 AS (
        SELECT membershipNumber,
            loyaltyStatus,
            totalMiles,
            NTILE(10) OVER (PARTITION BY loyaltyStatus ORDER BY totalMiles DESC) AS decileBucket
        FROM summed_miles_6
    ),
    map_1 AS (
        SELECT membershipNumber,
            MAP_FROM_ARRAYS(COLLECT_LIST(loyaltyStatus), COLLECT_LIST(decileBucket)) AS decileMonth1
        FROM rankings_1
        GROUP BY membershipNumber
    ),
    map_3 AS (
        SELECT membershipNumber,
            MAP_FROM_ARRAYS(COLLECT_LIST(loyaltyStatus), COLLECT_LIST(decileBucket)) AS decileMonth3
        FROM rankings_3
        GROUP BY membershipNumber
    ),
    map_6 AS (
        SELECT membershipNumber,
            MAP_FROM_ARRAYS(COLLECT_LIST(loyaltyStatus), COLLECT_LIST(decileBucket)) AS decileMonth6
        FROM rankings_6
        GROUP BY membershipNumber
    ),
    all_memberships AS (
        SELECT DISTINCT _profilefoundationreportingstg.membershipNumber AS membershipNumber FROM airline_loyalty_data
    )
    SELECT STRUCT(
            all_memberships.membershipNumber AS membershipNumber,
            STRUCT(
                    map_1.decileMonth1 AS decileMonth1,
                    map_3.decileMonth3 AS decileMonth3,
                    map_6.decileMonth6 AS decileMonth6
            ) AS decilesMileage
        ) AS _profilefoundationreportingstg
    FROM all_memberships
        LEFT JOIN map_1 ON  (all_memberships.membershipNumber = map_1.membershipNumber)
        LEFT JOIN map_3 ON  (all_memberships.membershipNumber = map_3.membershipNumber)
        LEFT JOIN map_6 ON  (all_memberships.membershipNumber = map_6.membershipNumber)
    }

summed_miles_1 AS (
    SELECT _profilefoundationreportingstg.membershipNumber AS membershipNumber,
           _profilefoundationreportingstg.loyaltyStatus AS loyaltyStatus,
           SUM(_profilefoundationreportingstg.mileage) AS totalMiles
    FROM airline_loyalty_data
    WHERE _profilefoundationreportingstg.transactionDate < (MAKE_DATE(YEAR(CURRENT_DATE), MONTH(CURRENT_DATE), 1) - MAKE_YM_INTERVAL(0, 0))
    GROUP BY 1,2
)

rankings_1 AS (
    SELECT membershipNumber,
           loyaltyStatus,
           totalMiles,
           NTILE(10) OVER (PARTITION BY loyaltyStatus ORDER BY totalMiles DESC) AS decileBucket
    FROM summed_miles_1
)

map_1 AS (
    SELECT membershipNumber,
           MAP_FROM_ARRAYS(COLLECT_LIST(loyaltyStatus), COLLECT_LIST(decileBucket)) AS decileMonth1
    FROM rankings_1
    GROUP BY membershipNumber
)

all_memberships AS (
    SELECT DISTINCT _profilefoundationreportingstg.membershipNumber AS membershipNumber FROM airline_loyalty_data
)

SELECT STRUCT(
           all_memberships.membershipNumber AS membershipNumber,
           STRUCT(
                map_1.decileMonth1 AS decileMonth1,
                map_3.decileMonth3 AS decileMonth3,
                map_6.decileMonth6 AS decileMonth6
           ) AS decilesMileage
       ) AS _profilefoundationreportingstg
FROM all_memberships
    LEFT JOIN map_1 ON  (all_memberships.membershipNumber = map_1.membershipNumber)
    LEFT JOIN map_3 ON  (all_memberships.membershipNumber = map_3.membershipNumber)
    LEFT JOIN map_6 ON  (all_memberships.membershipNumber = map_6.membershipNumber)

SELECT STRUCT(
           rankings.membershipNumber AS membershipNumber,
           STRUCT(
                MAP_FROM_ARRAYS(COLLECT_LIST(loyaltyStatus), COLLECT_LIST(decileBucket)) AS decileMonthAll
           ) AS decilesMileage
       ) AS _profilefoundationreportingstg
FROM rankings
GROUP BY rankings.membershipNumber