In [ ]:

    

// The code was removed by Watson Studio for sharing.


    

In [ ]:

    

dfData.createOrReplaceTempView("dfData")
dfData.count


    

In [ ]:

    

dfData.printSchema


    

In [ ]:

    

val results = spark.sql("SELECT dfData.doc.properties.naturecode FROM dfData")
results.collect


    

In [ ]:

    

val results2 = spark.sql("SELECT * FROM dfData WHERE dfData.doc.properties.naturecode='LARCEN'")
results2.collect


    

In [ ]:

    

results2.count