[SPARK-32762][SQL][TEST] Enhance the verification of ExpressionsSchemaSuite to sql-expression-schema.md

sql/core/src/test/resources/sql-functions/sql-expression-schema.md

@@ -1,6 +1,6 @@
 <!-- Automatically generated by ExpressionsSchemaSuite -->
 ## Summary
-  - Number of queries: 338
+  - Number of queries: 339
   - Number of expressions that missing example: 34
   - Expressions missing examples: and,string,tinyint,double,smallint,date,decimal,boolean,float,binary,bigint,int,timestamp,struct,cume_dist,dense_rank,input_file_block_length,input_file_block_start,input_file_name,lag,lead,monotonically_increasing_id,ntile,!,not,or,percent_rank,rank,row_number,spark_partition_id,version,window,positive,count_min_sketch
 ## Schema of Built-in Functions

sql/core/src/test/scala/org/apache/spark/sql/ExpressionsSchemaSuite.scala

@@ -152,23 +152,37 @@ class ExpressionsSchemaSuite extends QueryTest with SharedSparkSession {
 
     val outputSize = outputs.size
     val headerSize = header.size
-    val expectedOutputs: Seq[QueryOutput] = {
+    val (expectedMissingExamples, expectedOutputs) = {
       val expectedGoldenOutput = fileToString(resultFile)
       val lines = expectedGoldenOutput.split("\n")
       val expectedSize = lines.size
 
       assert(expectedSize == outputSize + headerSize,
         s"Expected $expectedSize blocks in result file but got " +
-          s"${outputSize + headerSize}. Try regenerate the result files.")
+          s"${outputSize + headerSize}. Try regenerating the result files.")
 
-      Seq.tabulate(outputSize) { i =>
+      val numberOfQueries = lines(2).split(":")(1).trim.toInt
+      val numberOfMissingExample = lines(3).split(":")(1).trim.toInt
+      val missingExamples = lines(4).split(":")(1).trim.split(",")
+      val expectedOutputs = Seq.tabulate(outputSize) { i =>
         val segments = lines(i + headerSize).split('|')
         QueryOutput(
           className = segments(1).trim,
           funcName = segments(2).trim,
           sql = segments(3).trim,
           schema = segments(4).trim)
       }
+
+      // Ensure consistency of the result file.
+      assert(numberOfQueries == expectedOutputs.size,
+        s"outputs size: ${expectedOutputs.size} not same as numberOfQueries: $numberOfQueries " +
+          "record in result file. Try regenerating the result files.")
+      assert(numberOfMissingExample == missingExamples.size,
+        s"missing examples size: ${missingExamples.size} not same as " +
+          s"numberOfMissingExample: $numberOfMissingExample " +
+          "record in result file. Try regenerating the result files.")
+
+      (missingExamples, expectedOutputs)
     }
 
     // Compare results.
@@ -179,5 +193,13 @@ class ExpressionsSchemaSuite extends QueryTest with SharedSparkSession {
       assert(expected.sql == output.sql, "SQL query did not match")
       assert(expected.schema == output.schema, s"Schema did not match for query ${expected.sql}")
     }
+
+    // Compare expressions missing examples
+    assert(expectedMissingExamples.length == missingExamples.size,
+      "The number of missing examples not equals the number of expected missing examples.")
+
+    missingExamples.zip(expectedMissingExamples).foreach { case (output, expected) =>
+      assert(expected == output, "Missing example expression not match")
+    }
   }
 }