Support WhenMatched::Delete in merge_insert

Author: jtuglu1

docs/src/rest.yaml

@@ -737,6 +737,13 @@ paths:
         required: true
         schema:
           type: string
+      - name: "when_matched_delete"
+        in: query
+        description: Delete all rows in target table where a match exists in source table
+        required: false
+        schema:
+          type: boolean
+          default: false
       - name: "when_matched_update_all"
         in: query
         description: Update all columns when rows match
@@ -786,6 +793,7 @@ paths:
         It passes in the `MergeInsertIntoTableRequest` information in the following way:
         - `id`: pass through path parameter of the same name
         - `on`: pass through query parameter of the same name
+        - `when_matched_delete`: pass through query parameter of the same name
         - `when_matched_update_all`: pass through query parameter of the same name
         - `when_matched_update_all_filt`: pass through query parameter of the same name
         - `when_not_matched_insert_all`: pass through query parameter of the same name
@@ -1938,6 +1946,10 @@ components:
         "on":
           description: Column name to use for matching rows (required)
           type: string
+        when_matched_delete:
+          description: Delete all rows in target table where a match exists in source table
+          type: boolean
+          default: false
         when_matched_update_all:
           description: Update all columns when rows match
           type: boolean

java/lance-jni/Cargo.lock

@@ -113,6 +113,7 @@ fn extract_when_matched<'local>(env: &mut JNIEnv<'local>, jparam: &JObject) -> R
             Some(expr) => Ok(WhenMatched::UpdateIf(expr)),
             None => Err(Error::input_error("No matched updated expr".to_string())),
         },
+        "Delete" => Ok(WhenMatched::Delete),
         _ => Err(Error::input_error(format!(
             "Illegal when_matched: {when_matched}",
         ))),

java/lance-jni/src/merge_insert.rs

@@ -66,6 +66,19 @@ public MergeInsertParams withMatchedDoNothing() {
     return this;
   }
 
+  /**
+   * Specify that when a row in the source table matches a row in the target table, the row in the
+   * target table is deleted.
+   *
+   * <p>This can be used to achieve "when matched delete" behavior.
+   *
+   * @return This MergeInsertParams instance
+   */
+  public MergeInsertParams withMatchedDelete() {
+    this.whenMatched = WhenMatched.Delete;
+    return this;
+  }
+
   /**
    * Specify that when a row in the source table matches a row in the target table and the
    * expression evaluates to true, the row in the target table is updated by the matched row from
@@ -284,6 +297,9 @@ public enum WhenMatched {
      * true.
      */
     UpdateIf,
+
+    /** The row is deleted from the target table. */
+    Delete
   }
 
   public enum WhenNotMatched {

java/src/main/java/com/lancedb/lance/merge/MergeInsertParams.java

@@ -120,6 +120,24 @@ public void testWhenMatchedUpdateIf() throws Exception {
     }
   }
 
+  @Test
+  public void testWhenMatchedDelete() throws Exception {
+    // Test delete matched target rows if expression is true
+
+    try (VectorSchemaRoot source = buildSource(testDataset.getSchema(), allocator)) {
+      try (ArrowArrayStream sourceStream = convertToStream(source, allocator)) {
+        MergeInsertResult result =
+            dataset.mergeInsert(
+                new MergeInsertParams(Collections.singletonList("id"))
+                    .withMatchedDelete()
+                    .withNotMatched(MergeInsertParams.WhenNotMatched.DoNothing),
+                sourceStream);
+
+        Assertions.assertEquals("{3=Person 3, 4=Person 4}", readAll(result.dataset()).toString());
+      }
+    }
+  }
+
   @Test
   public void testWhenNotMatchedBySourceDelete() throws Exception {
     // Test delete target rows which are not matched with source.

java/src/test/java/com/lancedb/lance/MergeInsertTest.java

@@ -159,6 +159,16 @@ def when_matched_update_all(
         """
         return super(MergeInsertBuilder, self).when_matched_update_all(condition)
 
+    def when_matched_delete(self) -> "MergeInsertBuilder":
+        """
+        Configure the operation to delete matched rows in the target table.
+
+        After this method is called, when the merge insert operation executes,
+        any rows that match both the source table and the target table will be
+        deleted.
+        """
+        return super(MergeInsertBuilder, self).when_matched_delete()
+
     def when_not_matched_insert_all(self) -> "MergeInsertBuilder":
         """
         Configure the operation to insert not matched rows

python/python/lance/dataset.py

@@ -177,6 +177,11 @@ impl MergeInsertBuilder {
         Ok(slf)
     }
 
+    pub fn when_matched_delete(mut slf: PyRefMut<Self>) -> PyResult<PyRefMut<Self>> {
+        slf.builder.when_matched(WhenMatched::Delete);
+        Ok(slf)
+    }
+
     pub fn when_not_matched_insert_all(mut slf: PyRefMut<Self>) -> PyResult<PyRefMut<Self>> {
         slf.builder.when_not_matched(WhenNotMatched::InsertAll);
         Ok(slf)