[SPARK-55978][SQL] Add TABLESAMPLE SYSTEM block sampling with DSv2 pushdown

docs/sql-ref-ansi-compliance.md

@@ -429,6 +429,7 @@ Below is a list of all the keywords in Spark SQL.
 |ATOMIC|non-reserved|non-reserved|non-reserved|
 |AUTHORIZATION|reserved|non-reserved|reserved|
 |BEGIN|non-reserved|non-reserved|non-reserved|
+|BERNOULLI|non-reserved|non-reserved|non-reserved|
 |BETWEEN|non-reserved|non-reserved|reserved|
 |BIGINT|non-reserved|non-reserved|reserved|
 |BINARY|non-reserved|non-reserved|reserved|
@@ -753,6 +754,7 @@ Below is a list of all the keywords in Spark SQL.
 |SUBSTR|non-reserved|non-reserved|non-reserved|
 |SUBSTRING|non-reserved|non-reserved|non-reserved|
 |SYNC|non-reserved|non-reserved|non-reserved|
+|SYSTEM|non-reserved|non-reserved|reserved|
 |SYSTEM_TIME|non-reserved|non-reserved|non-reserved|
 |SYSTEM_VERSION|non-reserved|non-reserved|non-reserved|
 |TABLE|reserved|non-reserved|reserved|

sql/api/src/main/antlr4/org/apache/spark/sql/catalyst/parser/SqlBaseLexer.g4

@@ -148,6 +148,7 @@ AT: 'AT';
 ATOMIC: 'ATOMIC';
 AUTHORIZATION: 'AUTHORIZATION';
 BEGIN: 'BEGIN';
+BERNOULLI: 'BERNOULLI';
 BETWEEN: 'BETWEEN';
 BIGINT: 'BIGINT';
 BINARY: 'BINARY';
@@ -471,6 +472,7 @@ STRUCT: 'STRUCT' {incComplexTypeLevelCounter();};
 SUBSTR: 'SUBSTR';
 SUBSTRING: 'SUBSTRING';
 SYNC: 'SYNC';
+SYSTEM: 'SYSTEM';
 SYSTEM_TIME: 'SYSTEM_TIME';
 SYSTEM_VERSION: 'SYSTEM_VERSION';
 TABLE: 'TABLE';

sql/api/src/main/antlr4/org/apache/spark/sql/catalyst/parser/SqlBaseParser.g4

@@ -1049,7 +1049,9 @@ joinCriteria
     ;
 
 sample
-    : TABLESAMPLE LEFT_PAREN sampleMethod? RIGHT_PAREN (REPEATABLE LEFT_PAREN seed=integerValue RIGHT_PAREN)?
+    : TABLESAMPLE (sampleType=(SYSTEM | BERNOULLI))?
+      LEFT_PAREN sampleMethod? RIGHT_PAREN
+      (REPEATABLE LEFT_PAREN seed=integerValue RIGHT_PAREN)?
     ;
 
 sampleMethod
@@ -1917,6 +1919,7 @@ ansiNonReserved
     | AT
     | ATOMIC
     | BEGIN
+    | BERNOULLI
     | BETWEEN
     | BIGINT
     | BINARY
@@ -2183,6 +2186,7 @@ ansiNonReserved
     | SUBSTR
     | SUBSTRING
     | SYNC
+    | SYSTEM
     | SYSTEM_TIME
     | SYSTEM_VERSION
     | TABLES
@@ -2287,6 +2291,7 @@ nonReserved
     | ATOMIC
     | AUTHORIZATION
     | BEGIN
+    | BERNOULLI
     | BETWEEN
     | BIGINT
     | BINARY
@@ -2599,6 +2604,7 @@ nonReserved
     | SUBSTR
     | SUBSTRING
     | SYNC
+    | SYSTEM
     | SYSTEM_TIME
     | SYSTEM_VERSION
     | TABLE

sql/catalyst/src/main/java/org/apache/spark/sql/connector/read/SupportsPushDownTableSample.java

@@ -36,4 +36,24 @@ boolean pushTableSample(
       double upperBound,
       boolean withReplacement,
       long seed);
+
+  /**
+   * Pushes down SAMPLE to the data source with sample method awareness.
+   * Data sources can override this to distinguish SYSTEM (block) from BERNOULLI (row) sampling.
+   * By default, rejects SYSTEM sampling for backward compatibility and delegates BERNOULLI to
+   *   the 4-parameter version.
+   */
+  default boolean pushTableSample(
+      double lowerBound,
+      double upperBound,
+      boolean withReplacement,
+      long seed,
+      boolean isSystemSampling) {
+    if (isSystemSampling) {
+      // If the data source hasn't overridden this method, it must have not added support
+      // for SYSTEM sampling. Don't apply sample pushdown.
+      return false;
+    }
+    return pushTableSample(lowerBound, upperBound, withReplacement, seed);
+  }
 }

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

@@ -530,7 +530,7 @@ object UnsupportedOperationChecker extends Logging {
           throwError("Sorting is not supported on streaming DataFrames/Datasets, unless it is on " +
             "aggregated DataFrame/Dataset in Complete output mode")
 
-        case Sample(_, _, _, _, child) if child.isStreaming =>
+        case Sample(_, _, _, _, child, _) if child.isStreaming =>
           throwError("Sampling is not supported on streaming DataFrames/Datasets")
 
         case Window(windowExpression, _, _, child, _) if child.isStreaming =>

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/Optimizer.scala

@@ -1288,7 +1288,7 @@ object CollapseProject extends Rule[LogicalPlan] with AliasHelper {
         limit.copy(child = p2.copy(projectList = newProjectList))
       case Project(l1, r @ Repartition(_, _, p @ Project(l2, _))) if isRenaming(l1, l2) =>
         r.copy(child = p.copy(projectList = buildCleanedProjectList(l1, p.projectList)))
-      case Project(l1, s @ Sample(_, _, _, _, p2 @ Project(l2, _))) if isRenaming(l1, l2) =>
+      case Project(l1, s @ Sample(_, _, _, _, p2 @ Project(l2, _), _)) if isRenaming(l1, l2) =>
         s.copy(child = p2.copy(projectList = buildCleanedProjectList(l1, p2.projectList)))
       case o => o
     }

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/parser/AstBuilder.scala

@@ -2250,10 +2250,14 @@ class AstBuilder extends DataTypeAstBuilder
    * - TABLESAMPLE(x ROWS): Sample the table down to the given number of rows.
    * - TABLESAMPLE(x PERCENT) [REPEATABLE (y)]: Sample the table down to the given percentage with
    * seed 'y'. Note that percentages are defined as a number between 0 and 100.
+   * - TABLESAMPLE SYSTEM(x PERCENT): Sample by data source dependent blocks or file splits.
    * - TABLESAMPLE(BUCKET x OUT OF y) [REPEATABLE (z)]: Sample the table down to a 'x' divided by
    * 'y' fraction with seed 'z'.
    */
   private def withSample(ctx: SampleContext, query: LogicalPlan): LogicalPlan = withOrigin(ctx) {
+    val isSystem = ctx.sampleType != null &&
+      ctx.sampleType.getType == SqlBaseParser.SYSTEM
+
     // Create a sampled plan if we need one.
     def sample(fraction: Double, seed: Long): Sample = {
       // The range of fraction accepted by Sample is [0, 1]. Because Hive's block sampling
@@ -2263,13 +2267,18 @@ class AstBuilder extends DataTypeAstBuilder
       validate(fraction >= 0.0 - eps && fraction <= 1.0 + eps,
         s"Sampling fraction ($fraction) must be on interval [0, 1]",
         ctx)
-      Sample(0.0, fraction, withReplacement = false, seed, query)
+      val method = if (isSystem) SampleMethod.System else SampleMethod.Bernoulli
+      Sample(0.0, fraction, withReplacement = false, seed, query, method)
     }
 
     if (ctx.sampleMethod() == null) {
       throw QueryParsingErrors.emptyInputForTableSampleError(ctx)
     }
 
+    if (isSystem && ctx.seed != null) {
+      operationNotAllowed("TABLESAMPLE SYSTEM does not support REPEATABLE", ctx)
+    }
+
     val seed = if (ctx.seed != null) {
       ctx.seed.getText.toLong
     } else {
@@ -2278,6 +2287,9 @@ class AstBuilder extends DataTypeAstBuilder
 
     ctx.sampleMethod() match {
       case ctx: SampleByRowsContext =>
+        if (isSystem) {
+          operationNotAllowed("TABLESAMPLE SYSTEM only supports PERCENT sampling", ctx)
+        }
         Limit(expression(ctx.expression), query)
 
       case ctx: SampleByPercentileContext =>
@@ -2289,6 +2301,9 @@ class AstBuilder extends DataTypeAstBuilder
         sample(sign * fraction / 100.0d, seed)
 
       case ctx: SampleByBytesContext =>
+        if (isSystem) {
+          operationNotAllowed("TABLESAMPLE SYSTEM only supports PERCENT sampling", ctx)
+        }
         val bytesStr = ctx.bytes.getText
         if (bytesStr.matches("[0-9]+[bBkKmMgG]")) {
           throw QueryParsingErrors.tableSampleByBytesUnsupportedError("byteLengthLiteral", ctx)
@@ -2297,6 +2312,9 @@ class AstBuilder extends DataTypeAstBuilder
         }
 
       case ctx: SampleByBucketContext if ctx.ON() != null =>
+        if (isSystem) {
+          operationNotAllowed("TABLESAMPLE SYSTEM only supports PERCENT sampling", ctx)
+        }
         if (ctx.identifier != null) {
           throw QueryParsingErrors.tableSampleByBytesUnsupportedError(
             "BUCKET x OUT OF y ON colname", ctx)
@@ -2306,6 +2324,9 @@ class AstBuilder extends DataTypeAstBuilder
         }
 
       case ctx: SampleByBucketContext =>
+        if (isSystem) {
+          operationNotAllowed("TABLESAMPLE SYSTEM only supports PERCENT sampling", ctx)
+        }
         sample(ctx.numerator.getText.toDouble / ctx.denominator.getText.toDouble, seed)
     }
   }

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/basicLogicalOperators.scala

@@ -1912,6 +1912,14 @@ object SubqueryAlias {
   }
 }
 
+sealed trait SampleMethod extends Serializable
+object SampleMethod {
+  /** Row-level sampling (BERNOULLI). Each row independently selected. No I/O savings. */
+  case object Bernoulli extends SampleMethod
+  /** System-level sampling (SYSTEM). Entire partitions/splits included or skipped. */
+  case object System extends SampleMethod
+}
+
 /**
  * Sample the dataset.
  *
@@ -1921,13 +1929,15 @@ object SubqueryAlias {
  * @param withReplacement Whether to sample with replacement.
  * @param seed the random seed
  * @param child the LogicalPlan
+ * @param sampleMethod the sampling method (Bernoulli or System)
  */
 case class Sample(
     lowerBound: Double,
     upperBound: Double,
     withReplacement: Boolean,
     seed: Long,
-    child: LogicalPlan) extends UnaryNode {
+    child: LogicalPlan,
+    sampleMethod: SampleMethod = SampleMethod.Bernoulli) extends UnaryNode {
 
   val eps = RandomSampler.roundingEpsilon
   val fraction = upperBound - lowerBound

sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/parser/PlanParserSuite.scala

@@ -870,6 +870,160 @@ class PlanParserSuite extends AnalysisTest {
         stop = 65))
   }
 
+  test("SPARK-55978: TABLESAMPLE SYSTEM and BERNOULLI - basic parsing") {
+    val sql = "select * from t"
+    // SYSTEM produces SampleMethod.System
+    assertEqual(
+      s"$sql tablesample system (43 percent) as x",
+      Sample(0, .43d, withReplacement = false, 10L,
+        table("t").as("x"), SampleMethod.System).select(star()))
+    // BERNOULLI produces SampleMethod.Bernoulli
+    assertEqual(
+      s"$sql tablesample bernoulli (43 percent) as x",
+      Sample(0, .43d, withReplacement = false, 10L,
+        table("t").as("x"), SampleMethod.Bernoulli).select(star()))
+    // No qualifier defaults to Bernoulli (backward compat)
+    assertEqual(
+      s"$sql tablesample(43 percent) as x",
+      Sample(0, .43d, withReplacement = false, 10L,
+        table("t").as("x")).select(star()))
+  }
+
+  test("SPARK-55978: TABLESAMPLE SYSTEM - case insensitivity") {
+    val sql = "select * from t"
+    // Keywords are case-insensitive
+    assertEqual(
+      s"$sql TABLESAMPLE SYSTEM (43 PERCENT) as x",
+      Sample(0, .43d, withReplacement = false, 10L,
+        table("t").as("x"), SampleMethod.System).select(star()))
+    assertEqual(
+      s"$sql TabLeSaMpLe SyStEm (43 PeRcEnT) as x",
+      Sample(0, .43d, withReplacement = false, 10L,
+        table("t").as("x"), SampleMethod.System).select(star()))
+    assertEqual(
+      s"$sql TABLESAMPLE BERNOULLI (43 PERCENT) as x",
+      Sample(0, .43d, withReplacement = false, 10L,
+        table("t").as("x"), SampleMethod.Bernoulli).select(star()))
+  }
+
+  test("SPARK-55978: TABLESAMPLE SYSTEM - boundary fractions") {
+    val sql = "select * from t"
+    // 0 PERCENT
+    assertEqual(
+      s"$sql tablesample system (0 percent) as x",
+      Sample(0, 0d, withReplacement = false, 10L,
+        table("t").as("x"), SampleMethod.System).select(star()))
+    // 100 PERCENT
+    assertEqual(
+      s"$sql tablesample system (100 percent) as x",
+      Sample(0, 1d, withReplacement = false, 10L,
+        table("t").as("x"), SampleMethod.System).select(star()))
+    // Fractional percent
+    assertEqual(
+      s"$sql tablesample system (0.1 percent) as x",
+      Sample(0, 0.001d, withReplacement = false, 10L,
+        table("t").as("x"), SampleMethod.System).select(star()))
+  }
+
+  test("SPARK-55978: TABLESAMPLE SYSTEM - unsupported sample methods") {
+    val sql = "select * from t"
+    // SYSTEM + ROWS -> error
+    checkError(
+      exception = parseException(s"$sql tablesample system (100 rows)"),
+      condition = "_LEGACY_ERROR_TEMP_0035",
+      parameters = Map(
+        "message" -> "TABLESAMPLE SYSTEM only supports PERCENT sampling"),
+      context = ExpectedContext(
+        fragment = "tablesample system (100 rows)",
+        start = 16,
+        stop = 44))
+    // SYSTEM + BYTES -> error
+    checkError(
+      exception = parseException(s"$sql tablesample system (300M)"),
+      condition = "_LEGACY_ERROR_TEMP_0035",
+      parameters = Map(
+        "message" -> "TABLESAMPLE SYSTEM only supports PERCENT sampling"),
+      context = ExpectedContext(
+        fragment = "tablesample system (300M)",
+        start = 16,
+        stop = 40))
+    // SYSTEM + BUCKET -> error
+    checkError(
+      exception = parseException(s"$sql tablesample system (bucket 4 out of 10)"),
+      condition = "_LEGACY_ERROR_TEMP_0035",
+      parameters = Map(
+        "message" -> "TABLESAMPLE SYSTEM only supports PERCENT sampling"),
+      context = ExpectedContext(
+        fragment = "tablesample system (bucket 4 out of 10)",
+        start = 16,
+        stop = 54))
+  }
+
+  test("SPARK-55978: TABLESAMPLE BERNOULLI - REPEATABLE is supported") {
+    assertEqual(
+      "select * from t tablesample bernoulli (43 percent) repeatable (123)",
+      Sample(0, .43d, withReplacement = false, 123L,
+        table("t"), SampleMethod.Bernoulli).select(star()))
+  }
+
+  test("SPARK-55978: TABLESAMPLE SYSTEM - REPEATABLE not supported") {
+    val sql = "select * from t"
+    checkError(
+      exception = parseException(s"$sql tablesample system (43 percent) repeatable (123)"),
+      condition = "_LEGACY_ERROR_TEMP_0035",
+      parameters = Map(
+        "message" -> "TABLESAMPLE SYSTEM does not support REPEATABLE"),
+      context = ExpectedContext(
+        fragment = "tablesample system (43 percent) repeatable (123)",
+        start = 16,
+        stop = 63))
+  }
+
+  test("SPARK-55978: TABLESAMPLE SYSTEM - fraction out of range") {
+    val sql = "select * from t"
+    // > 100 PERCENT
+    checkError(
+      exception = parseException(s"$sql tablesample system (150 percent) as x"),
+      condition = "_LEGACY_ERROR_TEMP_0064",
+      parameters = Map("msg" -> "Sampling fraction (1.5) must be on interval [0, 1]"),
+      context = ExpectedContext(
+        fragment = "tablesample system (150 percent)",
+        start = 16,
+        stop = 47))
+    // Negative PERCENT
+    checkError(
+      exception = parseException(s"$sql tablesample system (-10 percent) as x"),
+      condition = "_LEGACY_ERROR_TEMP_0064",
+      parameters = Map("msg" -> "Sampling fraction (-0.1) must be on interval [0, 1]"),
+      context = ExpectedContext(
+        fragment = "tablesample system (-10 percent)",
+        start = 16,
+        stop = 47))
+  }
+
+  test("SPARK-55978: TABLESAMPLE SYSTEM and BERNOULLI as identifiers") {
+    // SYSTEM usable as column name (nonReserved)
+    assertEqual("SELECT system FROM t",
+      table("t").select($"system"))
+    // BERNOULLI usable as column name
+    assertEqual("SELECT bernoulli FROM t",
+      table("t").select($"bernoulli"))
+    // Usable as table alias
+    assertEqual("SELECT * FROM t system",
+      table("t").as("system").select(star()))
+    assertEqual("SELECT * FROM t bernoulli",
+      table("t").as("bernoulli").select(star()))
+  }
+
+  test("SPARK-55978: TABLESAMPLE SYSTEM - subquery and join contexts") {
+    // SYSTEM sample in subquery
+    assertEqual(
+      "SELECT * FROM (SELECT * FROM t TABLESAMPLE SYSTEM (50 PERCENT)) sub",
+      Sample(0, .5d, withReplacement = false, 10L,
+        table("t"), SampleMethod.System)
+        .select(star()).as("sub").select(star()))
+  }
+
   test("sub-query") {
     val plan = table("t0").select($"id")
     assertEqual("select id from (t0)", plan)

sql/connect/client/jdbc/src/test/scala/org/apache/spark/sql/connect/client/jdbc/SparkConnectDatabaseMetaDataSuite.scala

@@ -209,7 +209,7 @@ class SparkConnectDatabaseMetaDataSuite extends ConnectFunSuite with RemoteSpark
     withConnection { conn =>
       val metadata = conn.getMetaData
       // scalastyle:off line.size.limit
-      assert(metadata.getSQLKeywords === "ADD,AFTER,AGGREGATE,ALWAYS,ANALYZE,ANTI,ANY_VALUE,ARCHIVE,ASC,BINDING,BUCKET,BUCKETS,BYTE,CACHE,CASCADE,CATALOG,CATALOGS,CHANGE,CHANGES,CLEAR,CLUSTER,CLUSTERED,CODEGEN,COLLATION,COLLECTION,COLUMNS,COMMENT,COMPACT,COMPACTIONS,COMPENSATION,COMPUTE,CONCATENATE,CONTAINS,CONTINUE,COST,DATA,DATABASE,DATABASES,DATEADD,DATEDIFF,DATE_ADD,DATE_DIFF,DAYOFYEAR,DAYS,DBPROPERTIES,DEFINED,DEFINER,DELAY,DELIMITED,DESC,DFS,DIRECTORIES,DIRECTORY,DISTRIBUTE,DIV,DO,ELSEIF,ENFORCED,ESCAPED,EVOLUTION,EXCHANGE,EXCLUDE,EXCLUSIVE,EXIT,EXPLAIN,EXPORT,EXTEND,EXTENDED,FIELDS,FILEFORMAT,FIRST,FLOW,FOLLOWING,FORMAT,FORMATTED,FOUND,FUNCTIONS,GENERATED,GEOGRAPHY,GEOMETRY,HANDLER,HOURS,IDENTIFIED,IDENTIFIER,IF,IGNORE,ILIKE,IMMEDIATE,INCLUDE,INCLUSIVE,INCREMENT,INDEX,INDEXES,INPATH,INPUT,INPUTFORMAT,INVOKER,ITEMS,ITERATE,JSON,KEY,KEYS,LAST,LAZY,LEAVE,LEVEL,LIMIT,LINES,LIST,LOAD,LOCATION,LOCK,LOCKS,LOGICAL,LONG,LOOP,MACRO,MAP,MATCHED,MATERIALIZED,MEASURE,METRICS,MICROSECOND,MICROSECONDS,MILLISECOND,MILLISECONDS,MINUS,MINUTES,MONTHS,MSCK,NAME,NAMESPACE,NAMESPACES,NANOSECOND,NANOSECONDS,NORELY,NULLS,OFFSET,OPTION,OPTIONS,OUTPUTFORMAT,OVERWRITE,PARTITIONED,PARTITIONS,PERCENT,PIVOT,PLACING,PRECEDING,PRINCIPALS,PROCEDURES,PROPERTIES,PURGE,QUARTER,QUERY,RECORDREADER,RECORDWRITER,RECOVER,RECURSION,REDUCE,REFRESH,RELY,RENAME,REPAIR,REPEAT,REPEATABLE,REPLACE,RESET,RESPECT,RESTRICT,ROLE,ROLES,SCHEMA,SCHEMAS,SECONDS,SECURITY,SEMI,SEPARATED,SERDE,SERDEPROPERTIES,SETS,SHORT,SHOW,SINGLE,SKEWED,SORT,SORTED,SOURCE,STATISTICS,STORED,STRATIFY,STREAM,STREAMING,STRING,STRUCT,SUBSTR,SYNC,SYSTEM_TIME,SYSTEM_VERSION,TABLES,TARGET,TBLPROPERTIES,TERMINATED,TIMEDIFF,TIMESTAMPADD,TIMESTAMPDIFF,TIMESTAMP_LTZ,TIMESTAMP_NTZ,TINYINT,TOUCH,TRANSACTION,TRANSACTIONS,TRANSFORM,TRUNCATE,TRY_CAST,TYPE,UNARCHIVE,UNBOUNDED,UNCACHE,UNLOCK,UNPIVOT,UNSET,UNTIL,USE,VAR,VARIABLE,VARIANT,VERSION,VIEW,VIEWS,VOID,WATERMARK,WEEK,WEEKS,WHILE,X,YEARS,ZONE")
+      assert(metadata.getSQLKeywords === "ADD,AFTER,AGGREGATE,ALWAYS,ANALYZE,ANTI,ANY_VALUE,ARCHIVE,ASC,BERNOULLI,BINDING,BUCKET,BUCKETS,BYTE,CACHE,CASCADE,CATALOG,CATALOGS,CHANGE,CHANGES,CLEAR,CLUSTER,CLUSTERED,CODEGEN,COLLATION,COLLECTION,COLUMNS,COMMENT,COMPACT,COMPACTIONS,COMPENSATION,COMPUTE,CONCATENATE,CONTAINS,CONTINUE,COST,DATA,DATABASE,DATABASES,DATEADD,DATEDIFF,DATE_ADD,DATE_DIFF,DAYOFYEAR,DAYS,DBPROPERTIES,DEFINED,DEFINER,DELAY,DELIMITED,DESC,DFS,DIRECTORIES,DIRECTORY,DISTRIBUTE,DIV,DO,ELSEIF,ENFORCED,ESCAPED,EVOLUTION,EXCHANGE,EXCLUDE,EXCLUSIVE,EXIT,EXPLAIN,EXPORT,EXTEND,EXTENDED,FIELDS,FILEFORMAT,FIRST,FLOW,FOLLOWING,FORMAT,FORMATTED,FOUND,FUNCTIONS,GENERATED,GEOGRAPHY,GEOMETRY,HANDLER,HOURS,IDENTIFIED,IDENTIFIER,IF,IGNORE,ILIKE,IMMEDIATE,INCLUDE,INCLUSIVE,INCREMENT,INDEX,INDEXES,INPATH,INPUT,INPUTFORMAT,INVOKER,ITEMS,ITERATE,JSON,KEY,KEYS,LAST,LAZY,LEAVE,LEVEL,LIMIT,LINES,LIST,LOAD,LOCATION,LOCK,LOCKS,LOGICAL,LONG,LOOP,MACRO,MAP,MATCHED,MATERIALIZED,MEASURE,METRICS,MICROSECOND,MICROSECONDS,MILLISECOND,MILLISECONDS,MINUS,MINUTES,MONTHS,MSCK,NAME,NAMESPACE,NAMESPACES,NANOSECOND,NANOSECONDS,NORELY,NULLS,OFFSET,OPTION,OPTIONS,OUTPUTFORMAT,OVERWRITE,PARTITIONED,PARTITIONS,PERCENT,PIVOT,PLACING,PRECEDING,PRINCIPALS,PROCEDURES,PROPERTIES,PURGE,QUARTER,QUERY,RECORDREADER,RECORDWRITER,RECOVER,RECURSION,REDUCE,REFRESH,RELY,RENAME,REPAIR,REPEAT,REPEATABLE,REPLACE,RESET,RESPECT,RESTRICT,ROLE,ROLES,SCHEMA,SCHEMAS,SECONDS,SECURITY,SEMI,SEPARATED,SERDE,SERDEPROPERTIES,SETS,SHORT,SHOW,SINGLE,SKEWED,SORT,SORTED,SOURCE,STATISTICS,STORED,STRATIFY,STREAM,STREAMING,STRING,STRUCT,SUBSTR,SYNC,SYSTEM_TIME,SYSTEM_VERSION,TABLES,TARGET,TBLPROPERTIES,TERMINATED,TIMEDIFF,TIMESTAMPADD,TIMESTAMPDIFF,TIMESTAMP_LTZ,TIMESTAMP_NTZ,TINYINT,TOUCH,TRANSACTION,TRANSACTIONS,TRANSFORM,TRUNCATE,TRY_CAST,TYPE,UNARCHIVE,UNBOUNDED,UNCACHE,UNLOCK,UNPIVOT,UNSET,UNTIL,USE,VAR,VARIABLE,VARIANT,VERSION,VIEW,VIEWS,VOID,WATERMARK,WEEK,WEEKS,WHILE,X,YEARS,ZONE")
       // scalastyle:on line.size.limit
     }
   }

sql/connect/common/src/test/resources/query-tests/explain-results/sample_fraction_seed.explain

@@ -1,2 +1,2 @@
-Sample 0.0, 0.43, false, 9890823
+Sample 0.0, 0.43, false, 9890823, Bernoulli
 +- LocalRelation <empty>, [id#0L, a#0, b#0]