feat: support dayname and monthname expressions

Add Comet support for the Spark 4.0+ dayname and monthname expressions by
routing them through the Arrow-direct codegen dispatcher in the spark-4.0
shim, so they run Spark's own generated code for exact parity including
locale handling.

Also add tests confirming that several datetime functions already execute in
Comet via Spark's rewrite rules, and update the expression support doc to
reflect this coverage:

- to_date / to_timestamp / to_timestamp_ntz / to_timestamp_ltz reduce to
  Cast (no format) or GetTimestamp (with format)
- make_timestamp_ntz / make_timestamp_ltz (6-argument form) reduce to
  MakeTimestamp
- current_date / current_timestamp / now / curdate are constant-folded to
  literals before Comet sees the plan

Author: andygrove

docs/source/contributor-guide/spark_expressions_support.md

@@ -407,10 +407,13 @@
 
 - [x] add_months
 - [x] convert_timezone
-- [ ] curdate
-- [ ] current_date
+- [x] curdate
+  - Alias of `current_date`; constant-folded to a literal by Spark's `ComputeCurrentTime` rule before Comet sees the plan.
+- [x] current_date
+  - Constant-folded to a literal by Spark's `ComputeCurrentTime` rule before Comet sees the plan.
 - [ ] current_time
-- [ ] current_timestamp
+- [x] current_timestamp
+  - Constant-folded to a literal by Spark's `ComputeCurrentTime` rule before Comet sees the plan.
 - [x] current_timezone
 - [x] date_add
 - [x] date_diff
@@ -423,7 +426,8 @@
 - [x] datediff
 - [x] datepart
 - [x] day
-- [ ] dayname
+- [x] dayname
+  - Spark 4.0+. Has no native lowering; routed through the codegen dispatcher (runs Spark's own `doGenCode`), gated by `spark.comet.exec.scalaUDF.codegen.enabled` (default true).
 - [x] dayofmonth
 - [x] dayofweek
 - [x] dayofyear
@@ -442,15 +446,18 @@
 - [ ] make_interval
 - [ ] make_time
 - [x] make_timestamp
-- [ ] make_timestamp_ltz
-- [ ] make_timestamp_ntz
+- [x] make_timestamp_ltz
+  - The 6-argument form rewrites to `MakeTimestamp` and runs via the codegen dispatcher. The 2-argument `(date, time)` form requires the Spark 4.1 TIME type and falls back.
+- [x] make_timestamp_ntz
+  - The 6-argument form rewrites to `MakeTimestamp` and runs via the codegen dispatcher. The 2-argument `(date, time)` form requires the Spark 4.1 TIME type and falls back.
 - [ ] make_ym_interval
 - [x] minute
 - [x] month
 - [ ] monthname
 - [x] months_between
 - [x] next_day
-- [ ] now
+- [x] now
+  - Alias of `current_timestamp`; constant-folded to a literal by Spark's `ComputeCurrentTime` rule before Comet sees the plan.
 - [x] quarter
 - [x] second
 - [ ] session_window
@@ -459,11 +466,15 @@
 - [x] timestamp_micros
 - [x] timestamp_millis
 - [x] timestamp_seconds
-- [ ] to_date
+- [x] to_date
+  - Rewrites to `Cast` (no format, native) or `Cast(GetTimestamp(...))` (with format, via the codegen dispatcher) before Comet sees the plan.
 - [ ] to_time
-- [ ] to_timestamp
-- [ ] to_timestamp_ltz
-- [ ] to_timestamp_ntz
+- [x] to_timestamp
+  - Rewrites to `Cast` (no format, native) or `GetTimestamp` (with format, via the codegen dispatcher) before Comet sees the plan.
+- [x] to_timestamp_ltz
+  - Rewrites to `to_timestamp` with `TimestampType`; same support as `to_timestamp`.
+- [x] to_timestamp_ntz
+  - Rewrites to `to_timestamp` with `TimestampNTZType`; same support as `to_timestamp`.
 - [x] to_unix_timestamp
 - [x] to_utc_timestamp
   - Spark 3.4.3 (audited 2026-05-12): identical to 3.5.8.

spark/src/main/spark-4.0/org/apache/comet/shims/CometExprShim.scala

@@ -31,7 +31,7 @@ import org.apache.spark.sql.types.{ArrayType, BinaryType, BooleanType, DataTypes
 import org.apache.comet.{CometConf, CometExplainInfo}
 import org.apache.comet.CometSparkSessionExtensions.withFallbackReason
 import org.apache.comet.expressions.{CometCast, CometEvalMode}
-import org.apache.comet.serde.{CommonStringExprs, Compatible, ExprOuterClass, Incompatible, SupportLevel}
+import org.apache.comet.serde.{CometScalaUDF, CommonStringExprs, Compatible, ExprOuterClass, Incompatible, SupportLevel}
 import org.apache.comet.serde.ExprOuterClass.{BinaryOutputStyle, Expr}
 import org.apache.comet.serde.QueryPlanSerde.{exprToProtoInternal, optExprWithFallbackReason, scalarFunctionExprToProto, scalarFunctionExprToProtoWithReturnType, supportedScalarSortElementType}
 
@@ -183,6 +183,13 @@ trait CometExprShim extends CommonStringExprs {
           optExprWithFallbackReason(mapSortExpr, ms, ms.child)
         }
 
+      // dayname / monthname (Spark 4.0+) have no native lowering. Route them through the
+      // Arrow-direct codegen dispatcher so they run Spark's own doGenCode for exact parity
+      // (including locale handling). Returns None and falls back cleanly when the dispatcher is
+      // disabled via spark.comet.exec.scalaUDF.codegen.enabled.
+      case _: DayName | _: MonthName =>
+        CometScalaUDF.emitJvmCodegenDispatch(expr, inputs, binding)
+
       case _ => None
     }
   }

spark/src/main/spark-4.1/org/apache/comet/shims/CometExprShim.scala

@@ -32,7 +32,7 @@ import org.apache.spark.sql.types.{ArrayType, BinaryType, BooleanType, DataTypes
 import org.apache.comet.{CometConf, CometExplainInfo}
 import org.apache.comet.CometSparkSessionExtensions.withFallbackReason
 import org.apache.comet.expressions.{CometCast, CometEvalMode}
-import org.apache.comet.serde.{CommonStringExprs, Compatible, ExprOuterClass, Incompatible, SupportLevel}
+import org.apache.comet.serde.{CometScalaUDF, CommonStringExprs, Compatible, ExprOuterClass, Incompatible, SupportLevel}
 import org.apache.comet.serde.ExprOuterClass.{BinaryOutputStyle, Expr}
 import org.apache.comet.serde.QueryPlanSerde.{exprToProtoInternal, optExprWithFallbackReason, scalarFunctionExprToProto, scalarFunctionExprToProtoWithReturnType, supportedScalarSortElementType}
 
@@ -214,6 +214,13 @@ trait CometExprShim extends CommonStringExprs {
           optExprWithFallbackReason(mapSortExpr, ms, ms.child)
         }
 
+      // dayname / monthname (Spark 4.0+) have no native lowering. Route them through the
+      // Arrow-direct codegen dispatcher so they run Spark's own doGenCode for exact parity
+      // (including locale handling). Returns None and falls back cleanly when the dispatcher is
+      // disabled via spark.comet.exec.scalaUDF.codegen.enabled.
+      case _: DayName | _: MonthName =>
+        CometScalaUDF.emitJvmCodegenDispatch(expr, inputs, binding)
+
       case _ => None
     }
   }

spark/src/main/spark-4.2/org/apache/comet/shims/CometExprShim.scala

@@ -32,7 +32,7 @@ import org.apache.spark.sql.types.{ArrayType, BinaryType, BooleanType, DataTypes
 import org.apache.comet.{CometConf, CometExplainInfo}
 import org.apache.comet.CometSparkSessionExtensions.withFallbackReason
 import org.apache.comet.expressions.{CometCast, CometEvalMode}
-import org.apache.comet.serde.{CommonStringExprs, Compatible, ExprOuterClass, Incompatible, SupportLevel}
+import org.apache.comet.serde.{CometScalaUDF, CommonStringExprs, Compatible, ExprOuterClass, Incompatible, SupportLevel}
 import org.apache.comet.serde.ExprOuterClass.{BinaryOutputStyle, Expr}
 import org.apache.comet.serde.QueryPlanSerde.{exprToProtoInternal, optExprWithFallbackReason, scalarFunctionExprToProto, scalarFunctionExprToProtoWithReturnType, supportedScalarSortElementType}
 
@@ -214,6 +214,13 @@ trait CometExprShim extends CommonStringExprs {
           optExprWithFallbackReason(mapSortExpr, ms, ms.child)
         }
 
+      // dayname / monthname (Spark 4.0+) have no native lowering. Route them through the
+      // Arrow-direct codegen dispatcher so they run Spark's own doGenCode for exact parity
+      // (including locale handling). Returns None and falls back cleanly when the dispatcher is
+      // disabled via spark.comet.exec.scalaUDF.codegen.enabled.
+      case _: DayName | _: MonthName =>
+        CometScalaUDF.emitJvmCodegenDispatch(expr, inputs, binding)
+
       case _ => None
     }
   }

spark/src/test/scala/org/apache/comet/CometExpressionSuite.scala

@@ -538,6 +538,70 @@ class CometExpressionSuite extends CometTestBase with AdaptiveSparkPlanHelper {
     }
   }
 
+  // The following tests cover datetime functions that are not registered under their own name in
+  // the Comet expression maps but are still handled: Spark rewrites them to supported expressions
+  // (Cast / GetTimestamp / MakeTimestamp) before Comet sees the plan, or the codegen dispatcher
+  // runs Spark's own generated code.
+
+  test("to_date / to_timestamp without format run natively") {
+    withParquetTable(
+      Seq(("2024-01-15", "2024-01-15 10:30:45"), ("2020-12-31", "2020-12-31 23:59:59")),
+      "tbl") {
+      // These rewrite to Cast(string -> date/timestamp), which Comet supports natively.
+      checkSparkAnswerAndOperator("SELECT to_date(_1) FROM tbl")
+      checkSparkAnswerAndOperator("SELECT to_timestamp(_2) FROM tbl")
+      checkSparkAnswerAndOperator("SELECT to_timestamp_ntz(_2) FROM tbl")
+      checkSparkAnswerAndOperator("SELECT to_timestamp_ltz(_2) FROM tbl")
+      checkSparkAnswerAndOperator("SELECT try_to_timestamp(_2) FROM tbl")
+    }
+  }
+
+  test("to_date / to_timestamp with format run via codegen dispatcher") {
+    withSQLConf(CometConf.COMET_SCALA_UDF_CODEGEN_ENABLED.key -> "true") {
+      withParquetTable(
+        Seq(("2024-01-15", "2024-01-15 10:30:45"), ("2020-12-31", "2020-12-31 23:59:59")),
+        "tbl") {
+        // With a format these rewrite to GetTimestamp, which is routed through the dispatcher.
+        checkSparkAnswerAndOperator("SELECT to_date(_1, 'yyyy-MM-dd') FROM tbl")
+        checkSparkAnswerAndOperator("SELECT to_timestamp(_2, 'yyyy-MM-dd HH:mm:ss') FROM tbl")
+      }
+    }
+  }
+
+  test("make_timestamp_ntz / make_timestamp_ltz run via codegen dispatcher") {
+    withSQLConf(CometConf.COMET_SCALA_UDF_CODEGEN_ENABLED.key -> "true") {
+      withParquetTable(Seq((2024, 1, 15, 10, 30, 45), (2020, 12, 31, 23, 59, 59)), "tbl") {
+        // The 6-argument forms rewrite to MakeTimestamp, which is routed through the dispatcher.
+        checkSparkAnswerAndOperator("SELECT make_timestamp_ntz(_1, _2, _3, _4, _5, _6) FROM tbl")
+        checkSparkAnswerAndOperator("SELECT make_timestamp_ltz(_1, _2, _3, _4, _5, _6) FROM tbl")
+      }
+    }
+  }
+
+  test("dayname / monthname run via codegen dispatcher") {
+    assume(isSpark40Plus)
+    withSQLConf(CometConf.COMET_SCALA_UDF_CODEGEN_ENABLED.key -> "true") {
+      withParquetTable(
+        Seq(Tuple1("2024-01-15"), Tuple1("2024-06-30"), Tuple1("2020-12-31")),
+        "tbl") {
+        checkSparkAnswerAndOperator("SELECT dayname(CAST(_1 AS DATE)) FROM tbl")
+        checkSparkAnswerAndOperator("SELECT monthname(CAST(_1 AS DATE)) FROM tbl")
+      }
+    }
+  }
+
+  test("current_date / current_timestamp / now constant-folded before Comet") {
+    Seq("current_date()", "current_timestamp()", "now()", "current_date", "current_timestamp")
+      .foreach { fn =>
+        val plan = spark.sql(s"SELECT $fn AS r").queryExecution.optimizedPlan
+        val folded = plan.expressions.exists {
+          case Alias(_: Literal, _) => true
+          case _ => false
+        }
+        assert(folded, s"expected '$fn' to be constant-folded to a Literal before Comet")
+      }
+  }
+
   test("hour on int96 timestamp column") {
     import testImplicits._