Introduction

Function list

SedonaSQL supports SQL/MM Part3 Spatial SQL Standard. It includes four kinds of SQL operators as follows. All these operators can be directly called through:

var myDataFrame = sparkSession.sql("YOUR_SQL")

Alternatively, expr and selectExpr can be used:

myDataFrame.withColumn("geometry", expr("ST_*")).selectExpr("ST_*")

• Constructor: Construct a Geometry given an input string or coordinates
  • Example: ST_GeomFromWKT (string). Create a Geometry from a WKT String.
  • Documentation: Here
• Function: Execute a function on the given column or columns
  • Example: ST_Distance (A, B). Given two Geometry A and B, return the Euclidean distance of A and B.
  • Documentation: Here
• Aggregate function: Return the aggregated value on the given column
  • Example: ST_Envelope_Aggr (Geometry column). Given a Geometry column, calculate the entire envelope boundary of this column.
  • Documentation: Here
• Predicate: Execute a logic judgement on the given columns and return true or false
  • Example: ST_Contains (A, B). Check if A fully contains B. Return "True" if yes, else return "False".
  • Documentation: Here

Sedona also provides an Adapter to convert SpatialRDD <-> DataFrame. Please read Adapter Scaladoc

SedonaSQL supports SparkSQL query optimizer, documentation is Here

Quick start

The detailed explanation is here Write a SQL/DataFrame application.

1. Add Sedona-core and Sedona-SQL into your project POM.xml or build.sbt
2. Declare your Spark Session

   sparkSession = SparkSession.builder().
         config("spark.serializer","org.apache.spark.serializer.KryoSerializer").
         config("spark.kryo.registrator", "org.apache.sedona.core.serde.SedonaKryoRegistrator").
         master("local[*]").appName("mySedonaSQLdemo").getOrCreate()
   

3. Add the following line after your SparkSession declaration:

   import org.apache.sedona.sql.utils.SedonaSQLRegistrator
   SedonaSQLRegistrator.registerAll(sparkSession)
   

---

Last update: September 24, 2021 03:54:31