当然，有可能。但只在聚类键上。这意味着我需要使用其他东西作为分区键或“bucket”。对于本例，我将假设营销活动是时间敏感的，通过使用“month”作为bucket（分区），我们将获得良好的分发和方便的查询。

CREATE TABLE stackoverflow.customertocampaign (
    campaign_month int,
    customer_id int,
    campaign_id int,
    customer_name text,
    PRIMARY KEY (campaign_month, customer_id, campaign_id)
);

现在，我可以了 INSERT 你的报告中描述的数据 CustomerToCampaign 变量。然后，这个查询工作：

aploetz@cqlsh:stackoverflow> SELECT campaign_month, customer_id, campaign_id
                             FROM customertocampaign WHERE campaign_month=202004 
                             AND (customer_id,campaign_id) = (1,2);

 campaign_month | customer_id | campaign_id
----------------+-------------+-------------
         202004 |           1 |           2

(1 rows)