import org.apache.spark.sql.types.StructType
val schema = new StructType()
.add($"id".long.copy(nullable = false))
.add($"name".string.copy(nullable = false))
import org.apache.spark.sql.catalyst.encoders.RowEncoder
val encoder = RowEncoder(schema)
scala> println(encoder.serializer(0).numberedTreeString)
00 validateexternaltype(getexternalrowfield(assertnotnull(input[0, org.apache.spark.sql.Row, true]), 0, id), LongType) AS id#1640L
01 +- validateexternaltype(getexternalrowfield(assertnotnull(input[0, org.apache.spark.sql.Row, true]), 0, id), LongType)
02 +- getexternalrowfield(assertnotnull(input[0, org.apache.spark.sql.Row, true]), 0, id)
03 +- assertnotnull(input[0, org.apache.spark.sql.Row, true])
04 +- input[0, org.apache.spark.sql.Row, true]
StaticInvoke Non-SQL Expression
StaticInvoke
is an expression with no SQL representation that represents a static method call in Scala or Java.
StaticInvoke
supports Java code generation (aka whole-stage codegen) to evaluate itself.
StaticInvoke
is created when:
-
ScalaReflection
is requested for the deserializer or serializer for a Scala type -
RowEncoder is requested for
deserializerFor
or serializer for a Scala type -
JavaTypeInference
is requested fordeserializerFor
orserializerFor
Note
|
StaticInvoke is similar to CallMethodViaReflection expression.
|
Creating StaticInvoke Instance
StaticInvoke
takes the following when created:
-
Name of the method to call on the static object
-
Optional expressions to pass as input arguments to the function
-
Flag to control whether to propagate
nulls
or not (enabled by default). If any of the arguments isnull
,null
is returned instead of calling the function