fix: when target is table, the sink downstream pk is not set (#19515)

e2e_test/sink/sink_into_table/parallelism.slt

@@ -8,16 +8,16 @@ statement ok
 SET STREAMING_PARALLELISM TO 2;
 
 statement ok
-create table t_simple (v1 int, v2 int);
+create table t_simple (v1 int, v2 int) append only;
 
 statement ok
-create table m_simple (v1 int primary key, v2 int);
+create table m_simple (v1 int, v2 int) append only;
 
 statement ok
 SET STREAMING_PARALLELISM TO 3;
 
 statement ok
-create sink s_simple_1 into m_simple as select v1, v2 from t_simple;
+create sink s_simple_1 into m_simple as select v1, v2 from t_simple with (type = 'append-only');
 
 query I
 select distinct parallelism from rw_fragment_parallelism where name in ('t_simple', 'm_simple', 's_simple_1');

src/frontend/planner_test/src/lib.rs

@@ -850,6 +850,7 @@ impl TestCase {
                     false,
                     None,
                     None,
+                    false,
                 ) {
                     Ok(sink_plan) => {
                         ret.sink_plan = Some(explain_plan(&sink_plan.into()));

src/frontend/planner_test/tests/testdata/input/sink_into_table.yaml

@@ -9,3 +9,15 @@
     explain create sink ss into t from s with (type = 'append-only');
   expected_outputs:
     - explain_output
+- sql: |
+    create table t1 (a int primary key, b int);
+    create table t2 (a int, b int primary key);
+    explain create sink s into t1 from t2;
+  expected_outputs:
+    - explain_output
+- sql: |
+    create table t1 (a int primary key, b int);
+    create table t2 (a int, b int primary key);
+    explain create sink s into t1 as select b from t2;
+  expected_outputs:
+    - explain_output

src/frontend/planner_test/tests/testdata/output/emit_on_window_close.yaml

@@ -206,7 +206,7 @@
     emit on window close
     WITH (connector = 'blackhole');
   explain_output: |
-    StreamSink { type: upsert, columns: [tm, foo, bar, t._row_id(hidden), lag, max, sum], pk: [t._row_id, t.bar] }
+    StreamSink { type: upsert, columns: [tm, foo, bar, t._row_id(hidden), lag, max, sum], downstream_pk: [] }
     └─StreamEowcOverWindow { window_functions: [first_value(t.foo) OVER(PARTITION BY t.bar ORDER BY t.tm ASC ROWS BETWEEN 2 PRECEDING AND 2 PRECEDING), max(t.foo) OVER(PARTITION BY t.bar ORDER BY t.tm ASC ROWS BETWEEN 1 PRECEDING AND 1 FOLLOWING), sum(t.foo) OVER(PARTITION BY t.bar ORDER BY t.tm ASC ROWS BETWEEN 2 PRECEDING AND CURRENT ROW EXCLUDE CURRENT ROW)] }
       └─StreamEowcSort { sort_column: t.tm }
         └─StreamExchange { dist: HashShard(t.bar) }

src/frontend/planner_test/tests/testdata/output/sink.yaml

@@ -9,7 +9,7 @@
       table.name='t1sink',
       type='upsert');
   explain_output: |
-    StreamSink { type: upsert, columns: [v1, v2, v3, v5, t1.v4(hidden)], pk: [t1.v3, t1.v4] }
+    StreamSink { type: upsert, columns: [v1, v2, v3, v5, t1.v4(hidden)], downstream_pk: [t1.v1, t1.v2] }
     └─StreamExchange { dist: HashShard(t1.v1, t1.v2) }
       └─StreamTableScan { table: t1, columns: [v1, v2, v3, v5, v4] }
 - id: create_upsert_jdbc_sink_with_downstream_pk2
@@ -22,7 +22,7 @@
       table.name='t1sink',
       type='upsert');
   explain_output: |
-    StreamSink { type: upsert, columns: [v1, v2, v3, v5], pk: [t1.v1, t1.v2] }
+    StreamSink { type: upsert, columns: [v1, v2, v3, v5], downstream_pk: [t1.v3, t1.v5] }
     └─StreamExchange { dist: HashShard(t1.v3, t1.v5) }
       └─StreamTableScan { table: t1, columns: [v1, v2, v3, v5] }
 - id: create_upsert_jdbc_sink_with_downstream_pk1
@@ -36,7 +36,7 @@
       type='upsert');
   explain_output: |+
     Fragment 0
-    StreamSink { type: upsert, columns: [v1, v2, v3, v5, t1.v4(hidden)], pk: [t1.v3, t1.v4] }
+    StreamSink { type: upsert, columns: [v1, v2, v3, v5, t1.v4(hidden)], downstream_pk: [t1.v1, t1.v2] }
     ├── tables: [ Sink: 0 ]
     ├── output: [ t1.v1, t1.v2, t1.v3, t1.v5, t1.v4 ]
     ├── stream key: [ t1.v3, t1.v4 ]
@@ -88,7 +88,7 @@
       type='upsert');
   explain_output: |+
     Fragment 0
-    StreamSink { type: upsert, columns: [v1, v2, v3, v5], pk: [t1.v1, t1.v2] }
+    StreamSink { type: upsert, columns: [v1, v2, v3, v5], downstream_pk: [t1.v3, t1.v5] }
     ├── tables: [ Sink: 0 ]
     ├── output: [ t1.v1, t1.v2, t1.v3, t1.v5 ]
     ├── stream key: [ t1.v1, t1.v2 ]
@@ -150,7 +150,7 @@
       primary_key='v1,v2'
     );
   explain_output: |
-    StreamSink { type: upsert, columns: [v1, v2, v3, v5, t1.v4(hidden)], pk: [t1.v3, t1.v4] }
+    StreamSink { type: upsert, columns: [v1, v2, v3, v5, t1.v4(hidden)], downstream_pk: [t1.v1, t1.v2] }
     └─StreamExchange { dist: HashShard(t1.v1, t1.v2) }
       └─StreamTableScan { table: t1, columns: [v1, v2, v3, v5, v4] }
 - id: downstream_pk_same_with_upstream
@@ -163,7 +163,7 @@
       primary_key='v2,v1'
     );
   explain_output: |
-    StreamSink { type: upsert, columns: [v2, v1, count], pk: [t1.v1, t1.v2] }
+    StreamSink { type: upsert, columns: [v2, v1, count], downstream_pk: [t1.v2, t1.v1] }
     └─StreamProject { exprs: [t1.v2, t1.v1, count] }
       └─StreamHashAgg { group_key: [t1.v1, t1.v2], aggs: [count] }
         └─StreamExchange { dist: HashShard(t1.v1, t1.v2) }
@@ -173,7 +173,7 @@
     create table t2 (a int, b int, watermark for b as b - 4) append only;
     explain create sink sk1 from t2 emit on window close with (connector='blackhole');
   explain_output: |
-    StreamSink { type: upsert, columns: [a, b, t2._row_id(hidden)], pk: [t2._row_id] }
+    StreamSink { type: upsert, columns: [a, b, t2._row_id(hidden)], downstream_pk: [] }
     └─StreamEowcSort { sort_column: t2.b }
       └─StreamTableScan { table: t2, columns: [a, b, _row_id] }
 - id: create_mock_iceberg_sink_append_only_with_sparse_partition
@@ -236,7 +236,7 @@
       primary_key = 'v1'
     );
   explain_output: |
-    StreamSink { type: upsert, columns: [v1, v2, v3, v4, t1._row_id(hidden)], pk: [t1._row_id] }
+    StreamSink { type: upsert, columns: [v1, v2, v3, v4, t1._row_id(hidden)], downstream_pk: [t1.v1] }
     └─StreamExchange { dist: HashShard($expr1) }
       └─StreamProject { exprs: [t1.v1, t1.v2, t1.v3, t1.v4, t1._row_id, Row(t1.v1, IcebergTransform('bucket[1]':Varchar, t1.v2), IcebergTransform('truncate[1]':Varchar, t1.v3), null:Int32) as $expr1] }
         └─StreamTableScan { table: t1, columns: [v1, v2, v3, v4, _row_id] }
@@ -258,5 +258,5 @@
       primary_key = 'v1'
     );
   explain_output: |
-    StreamSink { type: upsert, columns: [v1, v2, v3, v4, t1._row_id(hidden)], pk: [t1._row_id] }
+    StreamSink { type: upsert, columns: [v1, v2, v3, v4, t1._row_id(hidden)], downstream_pk: [t1.v1] }
     └─StreamTableScan { table: t1, columns: [v1, v2, v3, v4, _row_id] }

src/frontend/planner_test/tests/testdata/output/sink_into_table.yaml

@@ -12,3 +12,20 @@
     StreamProject { exprs: [s.x, Proctime as $expr1, (Proctime - '00:01:00':Interval) as $expr2, null:Serial], output_watermarks: [$expr1, $expr2] }
     └─StreamSink { type: append-only, columns: [x, s._row_id(hidden)] }
       └─StreamTableScan { table: s, columns: [x, _row_id] }
+- sql: |
+    create table t1 (a int primary key, b int);
+    create table t2 (a int, b int primary key);
+    explain create sink s into t1 from t2;
+  explain_output: |
+    StreamProject { exprs: [t2.a, t2.b] }
+    └─StreamSink { type: upsert, columns: [a, b], downstream_pk: [t2.a] }
+      └─StreamExchange { dist: HashShard(t2.a) }
+        └─StreamTableScan { table: t2, columns: [a, b] }
+- sql: |
+    create table t1 (a int primary key, b int);
+    create table t2 (a int, b int primary key);
+    explain create sink s into t1 as select b from t2;
+  explain_output: |
+    StreamProject { exprs: [t2.b, null:Int32] }
+    └─StreamSink { type: upsert, columns: [b], downstream_pk: [t2.b] }
+      └─StreamTableScan { table: t2, columns: [b] }

src/frontend/src/handler/create_sink.rs

@@ -27,8 +27,7 @@ use risingwave_common::catalog::{
 };
 use risingwave_common::secret::LocalSecretManager;
 use risingwave_common::types::DataType;
-use risingwave_common::{bail, catalog};
-use risingwave_connector::sink::catalog::{SinkCatalog, SinkFormatDesc, SinkType};
+use risingwave_connector::sink::catalog::{SinkCatalog, SinkFormatDesc};
 use risingwave_connector::sink::iceberg::{IcebergConfig, ICEBERG_SINK};
 use risingwave_connector::sink::{
     CONNECTOR_TYPE_KEY, SINK_TYPE_OPTION, SINK_USER_FORCE_APPEND_ONLY_OPTION, SINK_WITHOUT_BACKFILL,
@@ -234,8 +233,6 @@ pub async fn gen_sink_plan(
         }
     }
 
-    let target_table = target_table_catalog.as_ref().map(|catalog| catalog.id());
-
     let sink_plan = plan_root.gen_sink_plan(
         sink_table_name,
         definition,
@@ -245,8 +242,9 @@ pub async fn gen_sink_plan(
         sink_from_table_name,
         format_desc,
         without_backfill,
-        target_table,
+        target_table_catalog.clone(),
         partition_info,
+        user_specified_columns,
     )?;
 
     let sink_desc = sink_plan.sink_desc().clone();
@@ -278,18 +276,7 @@ pub async fn gen_sink_plan(
             }
         }
 
-        let user_defined_primary_key_table =
-            !(table_catalog.append_only || table_catalog.row_id_index.is_some());
-
-        if !(user_defined_primary_key_table
-            || sink_catalog.sink_type == SinkType::AppendOnly
-            || sink_catalog.sink_type == SinkType::ForceAppendOnly)
-        {
-            return Err(RwError::from(ErrorCode::BindError(
-                "Only append-only sinks can sink to a table without primary keys.".to_string(),
-            )));
-        }
-
+        let table_columns_without_rw_timestamp = table_catalog.columns_without_rw_timestamp();
         let exprs = derive_default_column_project_for_sink(
             &sink_catalog,
             sink_plan.schema(),

src/frontend/src/optimizer/mod.rs

@@ -13,6 +13,7 @@
 // limitations under the License.
 use std::num::NonZeroU32;
 use std::ops::DerefMut;
+use std::sync::Arc;
 
 pub mod plan_node;
 
@@ -40,7 +41,7 @@ mod plan_expr_visitor;
 mod rule;
 
 use std::assert_matches::assert_matches;
-use std::collections::HashMap;
+use std::collections::{BTreeMap, HashMap};
 
 use fixedbitset::FixedBitSet;
 use itertools::Itertools as _;
@@ -50,7 +51,7 @@ use plan_expr_rewriter::ConstEvalRewriter;
 use property::Order;
 use risingwave_common::bail;
 use risingwave_common::catalog::{
-    ColumnCatalog, ColumnDesc, ColumnId, ConflictBehavior, Field, Schema, TableId,
+    ColumnCatalog, ColumnDesc, ColumnId, ConflictBehavior, Field, Schema,
 };
 use risingwave_common::types::DataType;
 use risingwave_common::util::column_index_mapping::ColIndexMapping;
@@ -82,6 +83,7 @@ use crate::optimizer::plan_node::{
 use crate::optimizer::plan_visitor::TemporalJoinValidator;
 use crate::optimizer::property::Distribution;
 use crate::utils::{ColIndexMappingRewriteExt, WithOptionsSecResolved};
+use crate::TableCatalog;
 
 /// `PlanRoot` is used to describe a plan. planner will construct a `PlanRoot` with `LogicalNode`.
 /// and required distribution and order. And `PlanRoot` can generate corresponding streaming or
@@ -939,8 +941,9 @@ impl PlanRoot {
         sink_from_table_name: String,
         format_desc: Option<SinkFormatDesc>,
         without_backfill: bool,
-        target_table: Option<TableId>,
+        target_table: Option<Arc<TableCatalog>>,
         partition_info: Option<PartitionComputeInfo>,
+        user_specified_columns: bool,
     ) -> Result<StreamSink> {
         let stream_scan_type = if without_backfill {
             StreamScanType::UpstreamOnly
@@ -958,12 +961,17 @@ impl PlanRoot {
             self.gen_optimized_stream_plan_inner(emit_on_window_close, stream_scan_type)?;
         assert_eq!(self.phase, PlanPhase::Stream);
         assert_eq!(stream_plan.convention(), Convention::Stream);
+        let target_columns_to_plan_mapping = target_table.as_ref().map(|t| {
+            let columns = t.columns_without_rw_timestamp();
+            self.target_columns_to_plan_mapping(&columns, user_specified_columns)
+        });
         StreamSink::create(
             stream_plan,
             sink_name,
             db_name,
             sink_from_table_name,
             target_table,
+            target_columns_to_plan_mapping,
             self.required_dist.clone(),
             self.required_order.clone(),
             self.out_fields.clone(),
@@ -993,6 +1001,39 @@ impl PlanRoot {
             .config()
             .streaming_use_snapshot_backfill()
     }
+
+    /// used when the plan has a target relation such as DML and sink into table, return the mapping from table's columns to the plan's schema
+    pub fn target_columns_to_plan_mapping(
+        &self,
+        tar_cols: &[ColumnCatalog],
+        user_specified_columns: bool,
+    ) -> Vec<Option<usize>> {
+        #[allow(clippy::disallowed_methods)]
+        let visible_cols: Vec<(usize, String)> = self
+            .out_fields
+            .ones()
+            .zip_eq(self.out_names.iter().cloned())
+            .collect_vec();
+
+        let visible_col_idxes = visible_cols.iter().map(|(i, _)| *i).collect_vec();
+        let visible_col_idxes_by_name = visible_cols
+            .iter()
+            .map(|(i, name)| (name.as_ref(), *i))
+            .collect::<BTreeMap<_, _>>();
+
+        tar_cols
+            .iter()
+            .enumerate()
+            .filter(|(_, tar_col)| tar_col.can_dml())
+            .map(|(tar_i, tar_col)| {
+                if user_specified_columns {
+                    visible_col_idxes_by_name.get(tar_col.name()).cloned()
+                } else {
+                    (tar_i < visible_col_idxes.len()).then(|| visible_cols[tar_i].0)
+                }
+            })
+            .collect()
+    }
 }
 
 fn find_version_column_index(