fix: don't flatten fields during cdf read

crates/core/src/delta_datafusion/cdf/scan_utils.rs

@@ -92,9 +92,9 @@ pub fn create_partition_values<F: FileAction>(
     Ok(file_groups)
 }
 
-pub fn create_cdc_schema(mut schema_fields: Vec<Field>, include_type: bool) -> SchemaRef {
+pub fn create_cdc_schema(mut schema_fields: Vec<Arc<Field>>, include_type: bool) -> SchemaRef {
     if include_type {
-        schema_fields.push(Field::new(CHANGE_TYPE_COL, DataType::Utf8, true));
+        schema_fields.push(Field::new(CHANGE_TYPE_COL, DataType::Utf8, true).into());
     }
     Arc::new(Schema::new(schema_fields))
 }

crates/core/src/operations/load_cdf.rs

@@ -260,10 +260,10 @@ impl CdfLoadBuilder {
 
         let partition_values = self.snapshot.metadata().partition_columns.clone();
         let schema = self.snapshot.input_schema()?;
-        let schema_fields: Vec<Field> = self
+        let schema_fields: Vec<Arc<Field>> = self
             .snapshot
             .input_schema()?
-            .flattened_fields()
+            .fields()
             .into_iter()
             .filter(|f| !partition_values.contains(f.name()))
             .cloned()

python/tests/test_cdf.py

@@ -601,24 +601,30 @@ def test_write_overwrite_unpartitioned_cdf(tmp_path, sample_data: pa.Table):
         engine="rust",
         configuration={"delta.enableChangeDataFeed": "true"},
     )
-
-    # expected_data = (
-    #     ds.dataset(sample_data)
-    #     .to_table()
-    #     .append_column(
-    #         field_=pa.field("_change_type", pa.string(), nullable=True),
-    #         column=[["delete"] * 5],
-    #     )
-    # )
+    sort_values = [("_change_type", "ascending"), ("utf8", "ascending")]
+    expected_data = (
+        ds.dataset(pa.concat_tables([sample_data] * 3))
+        .to_table()
+        .append_column(
+            field_=pa.field("_change_type", pa.string(), nullable=True),
+            column=[["delete"] * 5 + ["insert"] * 10],
+        )
+    ).sort_by(sort_values)
 
     assert not os.path.exists(
         cdc_path
     ), "_change_data shouldn't exist since table was overwritten"
 
     ## TODO(ion): check if you see insert and deletes in commit version 1
 
-    # assert dt.load_cdf().read_all().drop_columns(['_commit_version', '_commit_timestamp']) == expected_data
-    # assert dt.to_pyarrow_table().sort_by([("utf8", "ascending")]) == sample_data
+    assert (
+        dt.load_cdf()
+        .read_all()
+        .drop_columns(["_commit_version", "_commit_timestamp"])
+        .sort_by(sort_values)
+        == expected_data
+    )
+    assert dt.to_pyarrow_table().sort_by([("utf8", "ascending")]) == sample_data
 
 
 def test_write_overwrite_partitioned_cdf(tmp_path, sample_data: pa.Table):
@@ -632,35 +638,42 @@ def test_write_overwrite_partitioned_cdf(tmp_path, sample_data: pa.Table):
         configuration={"delta.enableChangeDataFeed": "true"},
     )
 
+    batch2 = ds.dataset(sample_data).to_table(filter=(pc.field("int64") > 3))
+
     dt = DeltaTable(tmp_path)
     write_deltalake(
         dt,
-        data=ds.dataset(sample_data).to_table(filter=(pc.field("int64") > 3)),
+        data=batch2,
         engine="rust",
         mode="overwrite",
         predicate="int64 > 3",
         partition_by=["int64"],
         configuration={"delta.enableChangeDataFeed": "true"},
     )
 
-    # expected_data = (
-    #     ds.dataset(sample_data)
-    #     .to_table()
-    #     .append_column(
-    #         field_=pa.field("_change_type", pa.string(), nullable=False),
-    #         column=[["delete"] * 5],
-    #     )
-    # )
     table_schema = dt.schema().to_pyarrow()
     table_schema = table_schema.insert(
         len(table_schema), pa.field("_change_type", pa.string(), nullable=False)
     )
-    # cdc_data = pq.read_table(cdc_path, schema=table_schema)
+
+    sort_values = [("_change_type", "ascending"), ("utf8", "ascending")]
+
+    first_batch = sample_data.append_column(
+        field_=pa.field("_change_type", pa.string(), nullable=True),
+        column=[["insert"] * 5],
+    )
+
+    expected_data = pa.concat_tables([batch2] * 2).append_column(
+        field_=pa.field("_change_type", pa.string(), nullable=True),
+        column=[["delete", "insert"]],
+    )
 
     assert not os.path.exists(
         cdc_path
     ), "_change_data shouldn't exist since a specific partition was overwritten"
 
-    ## TODO(ion): check if you see insert and deletes in commit version 1
-    # assert cdc_data == expected_data
-    # assert dt.to_pyarrow_table().sort_by([("int64", "ascending")]) == sample_data
+    assert dt.load_cdf().read_all().drop_columns(
+        ["_commit_version", "_commit_timestamp"]
+    ).sort_by(sort_values).select(expected_data.column_names) == pa.concat_tables(
+        [first_batch, expected_data]
+    ).sort_by(sort_values)