Adress comments

Author: tanelk

sql/catalyst/src/main/java/org/apache/spark/sql/catalyst/expressions/ExpressionInfo.java

@@ -105,9 +105,8 @@ public ExpressionInfo(
             String deprecated) {
         assert name != null;
         assert arguments != null;
-        assert arguments.isEmpty() || arguments.startsWith("\n    Arguments:\n");
         assert examples != null;
-        assert examples.isEmpty() || examples.startsWith("\n    Examples:\n");
+        assert examples.isEmpty() || examples.contains("    Examples:");
         assert note != null;
         assert group != null;
         assert since != null;
@@ -143,9 +142,9 @@ public ExpressionInfo(
                 "got [" + group + "].");
         }
         if (!since.isEmpty()) {
-            if (!since.matches("[0-9]+\\.[0-9]+\\.[0-9]+")) {
+            if (Integer.parseInt(since.split("\\.")[0]) < 0) {
                 throw new IllegalArgumentException("'since' is malformed in the expression [" +
-                    this.name + "]. It should follow the MAJOR.MINOR.PATCH pattern; however, " +
+                    this.name + "]. It should not start with a negative number; however, " +
                     "got [" + since + "].");
             }
             this.extended += "\n    Since: " + since + "\n";

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/FunctionRegistry.scala

@@ -663,10 +663,10 @@ object FunctionRegistry {
     val clazz = scala.reflect.classTag[Cast].runtimeClass
     val usage = "_FUNC_(expr) - Casts the value `expr` to the target data type `_FUNC_`."
     val examples = s"""
-    Examples:
-      > SELECT _FUNC_($exampleIn);
-       $exampleOut
-  """
+      |    Examples:
+      |      > SELECT _FUNC_($exampleIn);
+      |       $exampleOut
+      |  """.stripMargin
     val expressionInfo = new ExpressionInfo(clazz.getCanonicalName, null, name, usage, "",
       examples, "", "", "2.0.1", "")
     (name, (expressionInfo, builder))

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/complexTypeCreator.scala

@@ -350,10 +350,10 @@ object CreateStruct {
       "_FUNC_(col1, col2, col3, ...) - Creates a struct with the given field values.",
       "",
       """
-    Examples:
-      > SELECT _FUNC_(1, 2, 3);
-       {"col1":1,"col2":2,"col3":3}
-  """,
+        |    Examples:
+        |      > SELECT _FUNC_(1, 2, 3);
+        |       {"col1":1,"col2":2,"col3":3}
+        |  """.stripMargin,
       "",
       "",
       "1.4.0",

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

@@ -1,8 +1,8 @@
 <!-- Automatically generated by ExpressionsSchemaSuite -->
 ## Summary
   - Number of queries: 339
-  - Number of expressions that missing example: 0
-  - Expressions missing examples: 
+  - Number of expressions that missing example: 1
+  - Expressions missing examples: window
 ## Schema of Built-in Functions
 | Class name | Function name or alias | Query example | Output schema |
 | ---------- | ---------------------- | ------------- | ------------- |
@@ -272,7 +272,7 @@
 | org.apache.spark.sql.catalyst.expressions.Subtract | - | SELECT 2 - 1 | struct<(2 - 1):int> |
 | org.apache.spark.sql.catalyst.expressions.Tan | tan | SELECT tan(0) | struct<TAN(CAST(0 AS DOUBLE)):double> |
 | org.apache.spark.sql.catalyst.expressions.Tanh | tanh | SELECT tanh(0) | struct<TANH(CAST(0 AS DOUBLE)):double> |
-| org.apache.spark.sql.catalyst.expressions.TimeWindow | window | SELECT window(timestamp('1970-01-01 12:00'), '10 minute', '5 minute') | struct<window:struct<start:timestamp,end:timestamp>> |
+| org.apache.spark.sql.catalyst.expressions.TimeWindow | window | N/A | N/A |
 | org.apache.spark.sql.catalyst.expressions.ToDegrees | degrees | SELECT degrees(3.141592653589793) | struct<DEGREES(CAST(3.141592653589793 AS DOUBLE)):double> |
 | org.apache.spark.sql.catalyst.expressions.ToRadians | radians | SELECT radians(180) | struct<RADIANS(CAST(180 AS DOUBLE)):double> |
 | org.apache.spark.sql.catalyst.expressions.ToUTCTimestamp | to_utc_timestamp | SELECT to_utc_timestamp('2016-08-31', 'Asia/Seoul') | struct<to_utc_timestamp(CAST(2016-08-31 AS TIMESTAMP), Asia/Seoul):timestamp> |

sql/core/src/test/scala/org/apache/spark/sql/expressions/ExpressionInfoSuite.scala

@@ -115,8 +115,14 @@ class ExpressionInfoSuite extends SparkFunSuite with SharedSparkSession {
       if (!ignoreSet.contains(info.getClassName)) {
         withClue(s"Function '${info.getName}', Expression class '${info.getClassName}'") {
           assert(info.getUsage.nonEmpty)
-          assert(info.getExamples.nonEmpty)
-          assert(info.getSince.nonEmpty)
+          assert(info.getExamples.startsWith("\n    Examples:\n"))
+          assert(info.getExamples.endsWith("\n  "))
+          assert(info.getSince.matches("[0-9]+\\.[0-9]+\\.[0-9]+"))
+
+          if (info.getArguments.nonEmpty) {
+            assert(info.getArguments.startsWith("\n    Arguments:\n"))
+            assert(info.getArguments.endsWith("\n  "))
+          }
         }
       }
     }
@@ -148,7 +154,12 @@ class ExpressionInfoSuite extends SparkFunSuite with SharedSparkSession {
       "org.apache.spark.sql.catalyst.expressions.Uuid",
       // The example calls methods that return unstable results.
       "org.apache.spark.sql.catalyst.expressions.CallMethodViaReflection",
-      "org.apache.spark.sql.catalyst.expressions.SparkVersion")
+      "org.apache.spark.sql.catalyst.expressions.SparkVersion",
+      "org.apache.spark.sql.catalyst.expressions.MonotonicallyIncreasingID",
+      "org.apache.spark.sql.catalyst.expressions.SparkPartitionID",
+      "org.apache.spark.sql.catalyst.expressions.InputFileName",
+      "org.apache.spark.sql.catalyst.expressions.InputFileBlockStart",
+      "org.apache.spark.sql.catalyst.expressions.InputFileBlockLength")
 
     val parFuncs = new ParVector(spark.sessionState.functionRegistry.listFunction().toVector)
     parFuncs.foreach { funcId =>