[SPARK-21336] Revise rand comparison in BatchEvalPythonExecSuite

## What changes were proposed in this pull request?

Revise rand comparison in BatchEvalPythonExecSuite

In BatchEvalPythonExecSuite, there are two cases using the case "rand() > 3"
Rand() generates a random value in [0, 1), it is wired to be compared with 3, use 0.3 instead

## How was this patch tested?

unit test

Please review http://spark.apache.org/contributing.html before opening a pull request.

Author: Wang Gengliang <ltnwgl@gmail.com>

Closes #18560 from gengliangwang/revise_BatchEvalPythonExecSuite.

sql/core/src/test/scala/org/apache/spark/sql/execution/python/BatchEvalPythonExecSuite.scala

@@ -65,7 +65,7 @@ class BatchEvalPythonExecSuite extends SparkPlanTest with SharedSQLContext {
 
   test("Python UDF: no push down on non-deterministic") {
     val df = Seq(("Hello", 4)).toDF("a", "b")
-      .where("b > 4 and dummyPythonUDF(a) and rand() > 3")
+      .where("b > 4 and dummyPythonUDF(a) and rand() > 0.3")
     val qualifiedPlanNodes = df.queryExecution.executedPlan.collect {
       case f @ FilterExec(
           And(_: AttributeReference, _: GreaterThan),
@@ -77,7 +77,7 @@ class BatchEvalPythonExecSuite extends SparkPlanTest with SharedSQLContext {
 
   test("Python UDF: no push down on predicates starting from the first non-deterministic") {
     val df = Seq(("Hello", 4)).toDF("a", "b")
-      .where("dummyPythonUDF(a) and rand() > 3 and b > 4")
+      .where("dummyPythonUDF(a) and rand() > 0.3 and b > 4")
     val qualifiedPlanNodes = df.queryExecution.executedPlan.collect {
       case f @ FilterExec(And(_: And, _: GreaterThan), InputAdapter(_: BatchEvalPythonExec)) => f
     }