SNOW-2381033: Deflake hybrid tests by using separate tables.

tests/integ/modin/hybrid/conftest.py

@@ -7,7 +7,7 @@
 import modin.pandas as pd
 from modin.config import context as config_context
 
-from tests.utils import IS_WINDOWS
+from tests.utils import IS_WINDOWS, Utils
 
 
 if IS_WINDOWS:
@@ -28,25 +28,35 @@ def enable_autoswitch():
 
 
 @pytest.fixture(scope="module")
-def init_transaction_tables():
+def module_scoped_test_table_name(session) -> str:
+    test_table_name = f"{Utils.random_table_name()}TESTTABLENAME"
+    try:
+        yield test_table_name
+    finally:
+        Utils.drop_table(session, test_table_name)
+
+
+@pytest.fixture(scope="module")
+def revenue_transactions(module_scoped_test_table_name):
     session = pd.session
     session.sql(
-        """
-    CREATE OR REPLACE TABLE revenue_transactions (
+        f"""
+    CREATE OR REPLACE TEMP TABLE {module_scoped_test_table_name} (
         Transaction_ID STRING,
         Date DATE,
         Revenue FLOAT
     );"""
     ).collect()
     session.sql(
-        """INSERT INTO revenue_transactions (Transaction_ID, Date, Revenue)
+        f"""INSERT INTO {module_scoped_test_table_name} (Transaction_ID, Date, Revenue)
     SELECT
         UUID_STRING() AS Transaction_ID,
         DATEADD(DAY, UNIFORM(0, 800, RANDOM(0)), '2024-01-01') AS Date,
         UNIFORM(10, 1000, RANDOM(0)) AS Revenue
-    FROM TABLE(GENERATOR(ROWCOUNT => 10000000));
+    FROM TABLE(GENERATOR(ROWCOUNT => 1000000));
     """
     ).collect()
+    return module_scoped_test_table_name
 
 
 @pytest.fixture

tests/integ/modin/hybrid/test_switch_operations.py

@@ -156,8 +156,8 @@ def test_move_to_me_cost_with_incompatible_dtype(caplog):
 
 # Newer version of modin switches before the merge
 @sql_count_checker(query_count=2 if MODIN_IS_AT_LEAST_0_37_0 else 0)
-def test_merge(init_transaction_tables, us_holidays_data):
-    df_transactions = pd.read_snowflake("REVENUE_TRANSACTIONS")
+def test_merge(revenue_transactions, us_holidays_data):
+    df_transactions = pd.read_snowflake(revenue_transactions)
     df_us_holidays = pd.DataFrame(us_holidays_data, columns=["Holiday", "Date"])
     assert df_transactions.get_backend() == "Snowflake"
     assert df_us_holidays.get_backend() == "Pandas"
@@ -174,9 +174,9 @@ def test_merge(init_transaction_tables, us_holidays_data):
 
 
 @sql_count_checker(query_count=2)
-def test_filtered_data(init_transaction_tables):
+def test_filtered_data(revenue_transactions):
     # When data is filtered, the engine should change when it is sufficiently small.
-    df_transactions = pd.read_snowflake("REVENUE_TRANSACTIONS")
+    df_transactions = pd.read_snowflake(revenue_transactions)
     assert df_transactions.get_backend() == "Snowflake"
     # in-place operations that do not change the backend
     # TODO: the following will result in an align which will grow the
@@ -203,7 +203,7 @@ def test_filtered_data(init_transaction_tables):
     # The SQL here is functionatly the same as above
     # Unlike in previous iterations of hybrid this does *not* move the data immediately
     df_transactions_filter2 = pd.read_snowflake(
-        "SELECT Date, SUM(Revenue) AS REVENUE FROM revenue_transactions WHERE Date >= DATEADD( 'days', -7, '2025-06-09' ) and Date < '2025-06-09' GROUP BY DATE"
+        f"SELECT Date, SUM(Revenue) AS REVENUE FROM {revenue_transactions} WHERE Date >= DATEADD( 'days', -7, '2025-06-09' ) and Date < '2025-06-09' GROUP BY DATE"
     )
     # We do not know the size of this data yet, because the query is entirely lazy
     assert df_transactions_filter2.get_backend() == "Snowflake"
@@ -227,8 +227,8 @@ def test_filtered_data(init_transaction_tables):
 
 
 @sql_count_checker(query_count=3)
-def test_apply(init_transaction_tables, us_holidays_data):
-    df_transactions = pd.read_snowflake("REVENUE_TRANSACTIONS").head(1000)
+def test_apply(revenue_transactions, us_holidays_data):
+    df_transactions = pd.read_snowflake(revenue_transactions).head(1000)
     assert df_transactions.get_backend() == "Snowflake"
     df_us_holidays = pd.DataFrame(us_holidays_data, columns=["Holiday", "Date"])
     df_us_holidays["Date"] = pd.to_datetime(df_us_holidays["Date"])
@@ -344,9 +344,9 @@ def test_explain_switch_empty():
 
 # Newer version of modin switches before the merge
 @sql_count_checker(query_count=2 if MODIN_IS_AT_LEAST_0_37_0 else 0)
-def test_explain_switch(init_transaction_tables, us_holidays_data):
+def test_explain_switch(revenue_transactions, us_holidays_data):
     clear_hybrid_switch_log()
-    df_transactions = pd.read_snowflake("REVENUE_TRANSACTIONS")
+    df_transactions = pd.read_snowflake(revenue_transactions)
     df_us_holidays = pd.DataFrame(us_holidays_data, columns=["Holiday", "Date"])
     pd.merge(df_us_holidays, df_transactions, left_on="Date", right_on="DATE")
     assert "decision" in str(pd.explain_switch())
@@ -459,7 +459,7 @@ def test_to_datetime():
     high_count_expected=True,
     high_count_reason="tests queries across different execution modes",
 )
-def test_query_count_no_switch(init_transaction_tables, use_session_param):
+def test_query_count_no_switch(revenue_transactions, use_session_param):
     """
     Tests that when there is no switching behavior the query count is the
     same under hybrid mode and non-hybrid mode.
@@ -471,7 +471,7 @@ def inner_test(df_in):
         df_result["COUNT"] = df_result.groupby("DATE")["REVENUE"].transform("count")
         return df_result
 
-    df_transactions = pd.read_snowflake("REVENUE_TRANSACTIONS")
+    df_transactions = pd.read_snowflake(revenue_transactions)
     inner_test(df_transactions)
     orig_len = None
     hybrid_len = None