Spark SQL是否区分大小写？

Question

看起来spark sql对“喜欢”查询区分大小写，对吗？

spark.sql("select distinct status, length(status)  from table")

返回

Active|6

spark.sql("select distinct status  from table where status like '%active%'")

不返回任何值

spark.sql("select distinct status  from table where status like '%Active%'")

返回

 Active

Answer 1

是的，Spark区分大小写。默认情况下，大多数RDBMS区分大小写以进行字符串比较。如果您希望不区分大小写，请尝试rlike或将列转换为大写/小写。

scala> val df = Seq(("Active"),("Stable"),("Inactive")).toDF("status")
df: org.apache.spark.sql.DataFrame = [status: string]

scala> df.createOrReplaceTempView("tbl")

scala> df.show
+--------+
|  status|
+--------+
|  Active|
|  Stable|
|Inactive|
+--------+


scala> spark.sql(""" select status from tbl where status like '%Active%' """).show
+------+
|status|
+------+
|Active|
+------+


scala> spark.sql(""" select status from tbl where lower(status) like '%active%' """).show
+--------+
|  status|
+--------+
|  Active|
|Inactive|
+--------+


scala>