sql - 根据 id 匹配时另一个表中的值更新表中的列

Question

我有两个表:

activities (
    org_id bigint,
    user_id bigint,
    lang text,
    timestamp bigint,
    ...
    PRIMARY KEY ((org_id, user_id), timestamp)

和

metadata (
    user_id text,
    org_id text,
    lang text,
    date_str text,
    ...
    PRIMARY KEY ((user_id, org_id), date_str)

鉴于: 事件.行>>元数据.行

我需要

update metadata and set metadata.lang = activities.lang for all rows in metadata,

where metadata.user_id = activities.user_id and metadata.org_id = activities.org_id

问:什么是优雅的短 cql 查询来实现相同的目的？

我尝试过:

update metadata set metadata.lang = (select activities.lang from

activities where activities.user_id = metadata.user_id and activities.org_id = metadata.org_id)

这显然因违反基数而失败。

Answer 1

CQL 不支持此类更新语句(请参阅 docs )。您需要在 where 条件中提供实际值。

因此，您需要在代码中完成所有操作(获取数据、生成更新语句等)，或者您可以使用 Spark 之类的东西，如下所示(未测试):

metadata = spark.read\
    .format("org.apache.spark.sql.cassandra")\
    .options(table="metadata", keyspace="test")\
    .load()\
    .select("user_id", "org_id")
activities = spark.read\
    .format("org.apache.spark.sql.cassandra")\
    .options(table="activities", keyspace="test")\
    .load()\
    .select("user_id", "org_id", "lang")
joined = activities.join(metadata, 
   (metadata["user_id"] == activitites["user_id"]) &
   (metadata["org_id"] == activities["org_id"]), "inner")\
   .select(metadata["user_id"], metadata["org_id"], "lang")
joined.write\
    .format("org.apache.spark.sql.cassandra")\
    .mode('append')\
    .options(table="metadata", keyspace="test")\
    .save()