如何获取表1中存在但表2中不存在的记录?

Ant*_*ony 1 sql apache-spark apache-spark-sql

我有两个表:excluded 和 kaggleresults。我正在尝试查找存在excluded但不存在的记录kaggleresults

计数:

scala> spark.sql("select * from excluded").count()
res136: Long = 4652

scala> spark.sql("select * from kaggleresults").count()
res137: Long = 4635
Run Code Online (Sandbox Code Playgroud)

区别在于 17

scala> res136-res137
res139: Long = 17
Run Code Online (Sandbox Code Playgroud)

我正在尝试获取这 17 条记录。我写了下面的查询,但它返回38

scala> spark.sql("select * from excluded left join kaggleresults on kaggleresults.subject_id = excluded.subject_id where kaggleresults.subject_id is null").count()
res135: Long = 38
Run Code Online (Sandbox Code Playgroud)

我需要写什么查询来获取这 17 条记录?

Jac*_*ski 5

LEFT_ANTI join不就是这样吗?

scala> val excluded = (0 to 5).toDS
left: org.apache.spark.sql.Dataset[Int] = [value: int]

scala> val kaggleresults = (3 to 10).toDS
right: org.apache.spark.sql.Dataset[Int] = [value: int]

scala> excluded.join(kaggleresults, Seq("value"), "leftanti").show
+-----+
|value|
+-----+
|    0|
|    1|
|    2|
+-----+
Run Code Online (Sandbox Code Playgroud)