What's the difference between SparkSession.sql and Dataset.sqlContext.sql?

Question

I have the following snippets of the code and I wonder what is the difference between these two and which one should I use? I am using spark 2.2.

Dataset<Row> df = sparkSession.readStream()
    .format("kafka")
    .load();

df.createOrReplaceTempView("table");
df.printSchema();

Dataset<Row> resultSet =  df.sqlContext().sql("select value from table"); //sparkSession.sql(this.query);
StreamingQuery streamingQuery = resultSet
        .writeStream()
        .trigger(Trigger.ProcessingTime(1000))
        .format("console")
        .start();

vs

Dataset<Row> df = sparkSession.readStream()
    .format("kafka")
    .load();

df.createOrReplaceTempView("table");

Dataset<Row> resultSet =  sparkSession.sql("select value from table"); //sparkSession.sql(this.query);
StreamingQuery streamingQuery = resultSet
        .writeStream()
        .trigger(Trigger.ProcessingTime(1000))
        .format("console")
        .start();

Answer 1

There is a very subtle difference between sparkSession.sql("sql query") vs df.sqlContext().sql("sql query").

Please note that you can have zero, two or more SparkSessions in a single Spark application (but it's assumed you'll have at least and often only one SparkSession in a Spark SQL application).

Please also note that a Dataset is bound to the SparkSession it was created within and the SparkSession will never change.

You may be wondering why anyone would want it, but that gives you boundary between queries and you could use the same table names for different datasets and that is a very powerful feature of Spark SQL actually.

The following example shows the difference and hopefully will give you some idea why it's powerful after all.

scala> spark.version
res0: String = 2.3.0-SNAPSHOT

scala> :type spark
org.apache.spark.sql.SparkSession

scala> spark.sql("show tables").show
+--------+---------+-----------+
|database|tableName|isTemporary|
+--------+---------+-----------+
+--------+---------+-----------+

scala> val df = spark.range(5)
df: org.apache.spark.sql.Dataset[Long] = [id: bigint]

scala> df.sqlContext.sql("show tables").show
+--------+---------+-----------+
|database|tableName|isTemporary|
+--------+---------+-----------+
+--------+---------+-----------+

scala> val anotherSession = spark.newSession
anotherSession: org.apache.spark.sql.SparkSession = org.apache.spark.sql.SparkSession@195c5803

scala> anotherSession.range(10).createOrReplaceTempView("new_table")

scala> anotherSession.sql("show tables").show
+--------+---------+-----------+
|database|tableName|isTemporary|
+--------+---------+-----------+
|        |new_table|       true|
+--------+---------+-----------+


scala> df.sqlContext.sql("show tables").show
+--------+---------+-----------+
|database|tableName|isTemporary|
+--------+---------+-----------+
+--------+---------+-----------+