在数组中使用Spark SQL进行查询

Question

为了给出backfround，我使用

sqlContext.read.json(sn3://...)
df.registerTable("posts")

scala> posts.printSchema
root
 |-- command: string (nullable = true)
 |-- externalId: string (nullable = true)
 |-- sourceMap: struct (nullable = true)
 |    |-- hashtags: array (nullable = true)
 |    |    |-- element: string (containsNull = true)
 |    |-- url: string (nullable = true)
 |-- type: string (nullable = true)

sqlContext.sql("select sourceMap['hashtags'] as ht from posts where ht.contains('nike')");

Answer 1

我找到了参考Hive SQL的答案。

sqlContext.sql("select sourceMap['hashtags'] from posts where array_contains(sourceMap['hashtags'], 'nike')");

关键函数是array_contains()