Treat warning as error in CI/Dev (#973)

* Treat warning as error in CI/Dev

This will help us avoid propagating warnings to our users, as occurred in #971.

* fixup! Treat warning as error in CI/Dev

pyproject.toml

@@ -29,15 +29,18 @@ classifiers = [
   "Programming Language :: Python :: 3.8",
   "Programming Language :: Python :: 3.9",
   "Programming Language :: Python :: 3.10",
-  "Programming Language :: Python :: 3.11"
+  "Programming Language :: Python :: 3.11",
 ]
 packages = [
   { include = "pyiceberg" },
   { from = "vendor", include = "fb303" },
   { from = "vendor", include = "hive_metastore" },
   { include = "tests", format = "sdist" },
   { include = "Makefile", format = "sdist" },
-  { include = "NOTICE", format = ["sdist", "wheel"] }
+  { include = "NOTICE", format = [
+    "sdist",
+    "wheel",
+  ] },
 ]
 include = [
   { path = "dev", format = "sdist" },
@@ -62,8 +65,8 @@ pyarrow = { version = ">=9.0.0,<18.0.0", optional = true }
 pandas = { version = ">=1.0.0,<3.0.0", optional = true }
 duckdb = { version = ">=0.5.0,<2.0.0", optional = true }
 ray = [
-  { version = "==2.10.0", python = "<3.9", optional = true},
-  { version = ">=2.10.0,<3.0.0", python = ">=3.9", optional = true}
+  { version = "==2.10.0", python = "<3.9", optional = true },
+  { version = ">=2.10.0,<3.0.0", python = ">=3.9", optional = true },
 ]
 python-snappy = { version = ">=0.6.0,<1.0.0", optional = true }
 thrift = { version = ">=0.13.0,<1.0.0", optional = true }
@@ -599,13 +602,17 @@ markers = [
   "s3: marks a test as requiring access to s3 compliant storage (use with --aws-access-key-id, --aws-secret-access-key, and --endpoint args)",
   "adlfs: marks a test as requiring access to adlfs compliant storage (use with --adlfs.account-name, --adlfs.account-key, and --adlfs.endpoint args)",
   "integration: marks integration tests against Apache Spark",
-  "gcs: marks a test as requiring access to gcs compliant storage (use with --gs.token, --gs.project, and --gs.endpoint)"
+  "gcs: marks a test as requiring access to gcs compliant storage (use with --gs.token, --gs.project, and --gs.endpoint)",
 ]
 
 # Turns a warning into an error
-#filterwarnings = [
-#    "error"
-#]
+filterwarnings = [
+  "error",
+  "ignore:A plugin raised an exception during an old-style hookwrapper teardown.",
+  "ignore:unclosed <socket.socket",
+  # Remove this in a future release of PySpark.
+  "ignore:distutils Version classes are deprecated. Use packaging.version instead.",
+]
 
 [tool.black]
 line-length = 130

tests/catalog/test_sql.py

@@ -355,7 +355,7 @@ def test_write_pyarrow_schema(catalog: SqlCatalog, table_identifier: Identifier)
     namespace = Catalog.namespace_from(table_identifier_nocatalog)
     catalog.create_namespace(namespace)
     table = catalog.create_table(table_identifier, pyarrow_table.schema)
-    table.overwrite(pyarrow_table)
+    table.append(pyarrow_table)
 
 
 @pytest.mark.parametrize(

tests/integration/test_deletes.py

@@ -145,6 +145,7 @@ def test_rewrite_partitioned_table_with_null(spark: SparkSession, session_catalo
 
 @pytest.mark.integration
 @pytest.mark.parametrize("format_version", [1, 2])
+@pytest.mark.filterwarnings("ignore:Delete operation did not match any records")
 def test_partitioned_table_no_match(spark: SparkSession, session_catalog: RestCatalog, format_version: int) -> None:
     identifier = "default.table_partitioned_delete"
 
@@ -175,6 +176,7 @@ def test_partitioned_table_no_match(spark: SparkSession, session_catalog: RestCa
 
 
 @pytest.mark.integration
+@pytest.mark.filterwarnings("ignore:Merge on read is not yet supported, falling back to copy-on-write")
 def test_delete_partitioned_table_positional_deletes(spark: SparkSession, session_catalog: RestCatalog) -> None:
     identifier = "default.table_partitioned_delete"
 
@@ -223,6 +225,7 @@ def test_delete_partitioned_table_positional_deletes(spark: SparkSession, sessio
 
 
 @pytest.mark.integration
+@pytest.mark.filterwarnings("ignore:Merge on read is not yet supported, falling back to copy-on-write")
 def test_overwrite_partitioned_table(spark: SparkSession, session_catalog: RestCatalog) -> None:
     identifier = "default.table_partitioned_delete"
 
@@ -274,6 +277,7 @@ def test_overwrite_partitioned_table(spark: SparkSession, session_catalog: RestC
 
 
 @pytest.mark.integration
+@pytest.mark.filterwarnings("ignore:Merge on read is not yet supported, falling back to copy-on-write")
 def test_partitioned_table_positional_deletes_sequence_number(spark: SparkSession, session_catalog: RestCatalog) -> None:
     identifier = "default.table_partitioned_delete_sequence_number"
 

tests/integration/test_inspect_table.py

@@ -79,7 +79,7 @@ def test_inspect_snapshots(
     identifier = "default.table_metadata_snapshots"
     tbl = _create_table(session_catalog, identifier, properties={"format-version": format_version})
 
-    tbl.overwrite(arrow_table_with_null)
+    tbl.append(arrow_table_with_null)
     # should produce a DELETE entry
     tbl.overwrite(arrow_table_with_null)
     # Since we don't rewrite, this should produce a new manifest with an ADDED entry
@@ -295,7 +295,7 @@ def test_inspect_refs(
     tbl = _create_table(session_catalog, identifier, properties={"format-version": format_version})
 
     # write data to create snapshot
-    tbl.overwrite(arrow_table_with_null)
+    tbl.append(arrow_table_with_null)
 
     # create a test branch
     spark.sql(
@@ -667,7 +667,7 @@ def test_inspect_files(
 
     tbl = _create_table(session_catalog, identifier, properties={"format-version": format_version})
 
-    tbl.overwrite(arrow_table_with_null)
+    tbl.append(arrow_table_with_null)
 
     # append more data
     tbl.append(arrow_table_with_null)

tests/integration/test_writes/test_writes.py

@@ -256,7 +256,7 @@ def test_data_files(spark: SparkSession, session_catalog: Catalog, arrow_table_w
     identifier = "default.arrow_data_files"
     tbl = _create_table(session_catalog, identifier, {"format-version": "1"}, [])
 
-    tbl.overwrite(arrow_table_with_null)
+    tbl.append(arrow_table_with_null)
     # should produce a DELETE entry
     tbl.overwrite(arrow_table_with_null)
     # Since we don't rewrite, this should produce a new manifest with an ADDED entry
@@ -288,7 +288,7 @@ def get_current_snapshot_id(identifier: str) -> int:
             .snapshot_id
         )
 
-    tbl.overwrite(arrow_table_with_null)
+    tbl.append(arrow_table_with_null)
     assert tbl.current_snapshot().snapshot_id == get_current_snapshot_id(identifier)  # type: ignore
     tbl.overwrite(arrow_table_with_null)
     assert tbl.current_snapshot().snapshot_id == get_current_snapshot_id(identifier)  # type: ignore
@@ -330,7 +330,7 @@ def test_python_writes_special_character_column_with_spark_reads(
     arrow_table_with_special_character_column = pa.Table.from_pydict(TEST_DATA_WITH_SPECIAL_CHARACTER_COLUMN, schema=pa_schema)
     tbl = _create_table(session_catalog, identifier, {"format-version": format_version}, schema=pa_schema)
 
-    tbl.overwrite(arrow_table_with_special_character_column)
+    tbl.append(arrow_table_with_special_character_column)
     spark_df = spark.sql(f"SELECT * FROM {identifier}").toPandas()
     pyiceberg_df = tbl.scan().to_pandas()
     assert spark_df.equals(pyiceberg_df)
@@ -354,7 +354,7 @@ def test_python_writes_dictionary_encoded_column_with_spark_reads(
 
     tbl = _create_table(session_catalog, identifier, {"format-version": format_version}, schema=pa_schema)
 
-    tbl.overwrite(arrow_table)
+    tbl.append(arrow_table)
     spark_df = spark.sql(f"SELECT * FROM {identifier}").toPandas()
     pyiceberg_df = tbl.scan().to_pandas()
     assert spark_df.equals(pyiceberg_df)
@@ -393,7 +393,7 @@ def test_python_writes_with_small_and_large_types_spark_reads(
     arrow_table = pa.Table.from_pydict(TEST_DATA, schema=pa_schema)
     tbl = _create_table(session_catalog, identifier, {"format-version": format_version}, schema=pa_schema)
 
-    tbl.overwrite(arrow_table)
+    tbl.append(arrow_table)
     spark_df = spark.sql(f"SELECT * FROM {identifier}").toPandas()
     pyiceberg_df = tbl.scan().to_pandas()
     assert spark_df.equals(pyiceberg_df)
@@ -429,7 +429,7 @@ def get_data_files_count(identifier: str) -> int:
 
     # writes 1 data file since the table is smaller than default target file size
     assert arrow_table_with_null.nbytes < TableProperties.WRITE_TARGET_FILE_SIZE_BYTES_DEFAULT
-    tbl.overwrite(arrow_table_with_null)
+    tbl.append(arrow_table_with_null)
     assert get_data_files_count(identifier) == 1
 
     # writes 1 data file as long as table is smaller than default target file size
@@ -820,7 +820,7 @@ def test_inspect_snapshots(
     identifier = "default.table_metadata_snapshots"
     tbl = _create_table(session_catalog, identifier, properties={"format-version": format_version})
 
-    tbl.overwrite(arrow_table_with_null)
+    tbl.append(arrow_table_with_null)
     # should produce a DELETE entry
     tbl.overwrite(arrow_table_with_null)
     # Since we don't rewrite, this should produce a new manifest with an ADDED entry
@@ -979,6 +979,7 @@ def test_table_write_subset_of_schema(session_catalog: Catalog, arrow_table_with
 
 @pytest.mark.integration
 @pytest.mark.parametrize("format_version", [1, 2])
+@pytest.mark.filterwarnings("ignore:Delete operation did not match any records")
 def test_table_write_out_of_order_schema(session_catalog: Catalog, arrow_table_with_null: pa.Table, format_version: int) -> None:
     identifier = "default.test_table_write_out_of_order_schema"
     # rotate the schema fields by 1
@@ -989,6 +990,7 @@ def test_table_write_out_of_order_schema(session_catalog: Catalog, arrow_table_w
     tbl = _create_table(session_catalog, identifier, {"format-version": format_version}, schema=rotated_schema)
 
     tbl.overwrite(arrow_table_with_null)
+
     tbl.append(arrow_table_with_null)
     # overwrite and then append should produce twice the data
     assert len(tbl.scan().to_arrow()) == len(arrow_table_with_null) * 2