PolarDB这个条 SQL，强行列存的效果还没有行存查询高，怎么回事？

PolarDB这个条 SQL，强行列存的效果还没有行存查询高，怎么回事？SELECT
/+ SET_VAR(cost_threshold_for_imci=0) /
roi.merchant_id,
roi.sku_id,
roi.machine_code,
roi.door_num,
roi.sku_name,
SUM(
CASE
roi.create_type
WHEN 1 THEN
roi.sku_num
WHEN 3 THEN
roi.sku_num
WHEN 2 THEN

• roi.sku_num
  WHEN 5 THEN
  roi.sku_num
  WHEN 6 THEN
• roi.sku_num ELSE 0
  END
  ) AS skuSaleNum,
  SUM(
  CASE
  roi.create_type
  WHEN 1 THEN
  roi.amount
  WHEN 3 THEN
  roi.amount
  WHEN 2 THEN
  • roi.amount
    WHEN 5 THEN
    roi.amount
    WHEN 6 THEN
  • roi.amount ELSE 0
    END
    ) AS skuSaleAmount
    FROM
    r_order_item roi
    WHERE
    roi.order_id IN (
    SELECT
    r.order_id
    FROM
    r_order r
    WHERE
    r.anomaly_status IN ( 1, 9 )
    AND r.close_type IN ( 0, 1, 5, 9, 10 )
    AND r.company_id = 314
    AND r.merchant_id = 2532
    AND r.create_time BETWEEN '2024-04-01 22:07:13.389'
    AND '2024-04-18 23:59:59.999'
    AND r.amount != 0
    )
    AND roi.STATUS = 1
    AND roi.sku_id NOT IN ( 2, 10 )
    GROUP BY
    roi.machine_id,
    roi.door_num,
    roi.sku_id
    ORDER BY
    skuSaleNum DESC