FEAT-#6574: UserWarning no longer displayed when Series/DataFrames ar…

…e small (#7323)

Signed-off-by: Jayson Willey <91502167+Jayson729@users.noreply.github.com>

modin/pandas/dataframe.py

@@ -205,9 +205,6 @@ def __init__(
                 self._query_compiler = distributed_frame._query_compiler
                 return
 
-            warnings.warn(
-                "Distributing {} object. This may take some time.".format(type(data))
-            )
             if isinstance(data, pandas.Index):
                 pass
             elif (
@@ -253,6 +250,12 @@ def __init__(
             pandas_df = pandas.DataFrame(
                 data=data, index=index, columns=columns, dtype=dtype, copy=copy
             )
+            if pandas_df.size >= 1_000_000:
+                warnings.warn(
+                    "Distributing {} object. This may take some time.".format(
+                        type(data)
+                    )
+                )
             self._query_compiler = from_pandas(pandas_df)._query_compiler
         else:
             self._query_compiler = query_compiler

modin/pandas/series.py

@@ -137,26 +137,28 @@ def __init__(
                 query_compiler.columns = pandas.Index([MODIN_UNNAMED_SERIES_LABEL])
         if query_compiler is None:
             # Defaulting to pandas
-            warnings.warn(
-                "Distributing {} object. This may take some time.".format(type(data))
-            )
             if name is None:
                 name = MODIN_UNNAMED_SERIES_LABEL
                 if isinstance(data, pandas.Series) and data.name is not None:
                     name = data.name
 
-            query_compiler = from_pandas(
-                pandas.DataFrame(
-                    pandas.Series(
-                        data=data,
-                        index=index,
-                        dtype=dtype,
-                        name=name,
-                        copy=copy,
-                        fastpath=fastpath,
+            pandas_df = pandas.DataFrame(
+                pandas.Series(
+                    data=data,
+                    index=index,
+                    dtype=dtype,
+                    name=name,
+                    copy=copy,
+                    fastpath=fastpath,
+                )
+            )
+            if pandas_df.size >= 2_500_000:
+                warnings.warn(
+                    "Distributing {} object. This may take some time.".format(
+                        type(data)
                     )
                 )
-            )._query_compiler
+            query_compiler = from_pandas(pandas_df)._query_compiler
         self._query_compiler = query_compiler.columnarize()
         if name is not None:
             self.name = name

modin/tests/pandas/dataframe/test_default.py

@@ -1473,3 +1473,37 @@ def test_df_from_series_with_tuple_name():
     df_equals(pd.DataFrame(pandas.Series(name=("a", 1))), pandas_result)
     # 2. Creating a Modin DF from Modin Series
     df_equals(pd.DataFrame(pd.Series(name=("a", 1))), pandas_result)
+
+
+def test_large_df_warns_distributing_takes_time():
+    # https://github.com/modin-project/modin/issues/6574
+
+    regex = r"Distributing (.*) object\. This may take some time\."
+    with pytest.warns(UserWarning, match=regex):
+        pd.DataFrame(np.random.randint(1_000_000, size=(100_000, 10)))
+
+
+def test_large_series_warns_distributing_takes_time():
+    # https://github.com/modin-project/modin/issues/6574
+
+    regex = r"Distributing (.*) object\. This may take some time\."
+    with pytest.warns(UserWarning, match=regex):
+        pd.Series(np.random.randint(1_000_000, size=(2_500_000)))
+
+
+def test_df_does_not_warn_distributing_takes_time():
+    # https://github.com/modin-project/modin/issues/6574
+
+    regex = r"Distributing (.*) object\. This may take some time\."
+    with warnings.catch_warnings():
+        warnings.filterwarnings("error", regex, UserWarning)
+        pd.DataFrame(np.random.randint(1_000_000, size=(100_000, 9)))
+
+
+def test_series_does_not_warn_distributing_takes_time():
+    # https://github.com/modin-project/modin/issues/6574
+
+    regex = r"Distributing (.*) object\. This may take some time\."
+    with warnings.catch_warnings():
+        warnings.filterwarnings("error", regex, UserWarning)
+        pd.Series(np.random.randint(1_000_000, size=(2_400_000)))

modin/tests/pandas/test_series.py

@@ -18,6 +18,7 @@
 import json
 import sys
 import unittest.mock as mock
+import warnings
 
 import matplotlib
 import numpy as np
@@ -26,7 +27,7 @@
 import pytest
 from numpy.testing import assert_array_equal
 from pandas.core.indexing import IndexingError
-from pandas.errors import SpecificationError
+from pandas.errors import PerformanceWarning, SpecificationError
 
 import modin.pandas as pd
 from modin.config import Engine, NPartitions, StorageFormat
@@ -3429,13 +3430,10 @@ def test_sub(data):
 
 def test_6782():
     datetime_scalar = datetime.datetime(1970, 1, 1, 0, 0)
-    with pytest.warns(UserWarning) as warns:
-        _ = pd.Series([datetime.datetime(2000, 1, 1)]) - datetime_scalar
-        for warn in warns.list:
-            assert (
-                "Adding/subtracting object-dtype array to DatetimeArray not vectorized"
-                not in str(warn)
-            )
+    match = "Adding/subtracting object-dtype array to DatetimeArray not vectorized"
+    with warnings.catch_warnings():
+        warnings.filterwarnings("error", match, PerformanceWarning)
+        pd.Series([datetime.datetime(2000, 1, 1)]) - datetime_scalar
 
 
 @pytest.mark.parametrize("data", test_data_values, ids=test_data_keys)