fix: bpd.merge on common columns (#1905)

Author: chelsea-lin

bigframes/dataframe.py

@@ -3434,15 +3434,9 @@ def merge(
             )
             return DataFrame(result_block)
 
-        if on is None:
-            if left_on is None or right_on is None:
-                raise ValueError("Must specify `on` or `left_on` + `right_on`.")
-        else:
-            if left_on is not None or right_on is not None:
-                raise ValueError(
-                    "Can not pass both `on` and `left_on` + `right_on` params."
-                )
-            left_on, right_on = on, on
+        left_on, right_on = self._validate_left_right_on(
+            right, on, left_on=left_on, right_on=right_on
+        )
 
         if utils.is_list_like(left_on):
             left_on = list(left_on)  # type: ignore
@@ -3479,6 +3473,41 @@ def merge(
         )
         return DataFrame(block)
 
+    def _validate_left_right_on(
+        self,
+        right: DataFrame,
+        on: Union[blocks.Label, Sequence[blocks.Label], None] = None,
+        *,
+        left_on: Union[blocks.Label, Sequence[blocks.Label], None] = None,
+        right_on: Union[blocks.Label, Sequence[blocks.Label], None] = None,
+    ):
+        if on is not None:
+            if left_on is not None or right_on is not None:
+                raise ValueError(
+                    "Can not pass both `on` and `left_on` + `right_on` params."
+                )
+            return on, on
+
+        if left_on is not None and right_on is not None:
+            return left_on, right_on
+
+        left_cols = self.columns
+        right_cols = right.columns
+        common_cols = left_cols.intersection(right_cols)
+        if len(common_cols) == 0:
+            raise ValueError(
+                "No common columns to perform merge on."
+                f"Merge options: left_on={left_on}, "
+                f"right_on={right_on}, "
+            )
+        if (
+            not left_cols.join(common_cols, how="inner").is_unique
+            or not right_cols.join(common_cols, how="inner").is_unique
+        ):
+            raise ValueError(f"Data columns not unique: {repr(common_cols)}")
+
+        return common_cols, common_cols
+
     def join(
         self,
         other: Union[DataFrame, bigframes.series.Series],

tests/system/small/test_pandas.py

@@ -13,6 +13,7 @@
 # limitations under the License.
 
 from datetime import datetime
+import re
 import typing
 
 import pandas as pd
@@ -343,7 +344,7 @@ def test_merge_left_on_right_on(scalars_dfs, merge_how):
     assert_pandas_df_equal(bf_result, pd_result, ignore_order=True)
 
 
-def test_pd_merge_cross(scalars_dfs):
+def test_merge_cross(scalars_dfs):
     scalars_df, scalars_pandas_df = scalars_dfs
     left_columns = ["int64_col", "float64_col", "int64_too"]
     right_columns = ["int64_col", "bool_col", "string_col", "rowindex_2"]
@@ -398,6 +399,61 @@ def test_merge_series(scalars_dfs, merge_how):
     assert_pandas_df_equal(bf_result, pd_result, ignore_order=True)
 
 
+def test_merge_w_common_columns(scalars_dfs):
+    scalars_df, scalars_pandas_df = scalars_dfs
+    left_columns = ["int64_col", "int64_too"]
+    right_columns = ["int64_col", "bool_col"]
+
+    df = bpd.merge(
+        scalars_df[left_columns], scalars_df[right_columns], "inner", sort=True
+    )
+
+    pd_result = pd.merge(
+        scalars_pandas_df[left_columns],
+        scalars_pandas_df[right_columns],
+        "inner",
+        sort=True,
+    )
+    assert_pandas_df_equal(df.to_pandas(), pd_result, ignore_order=True)
+
+
+def test_merge_raises_error_when_no_common_columns(scalars_dfs):
+    scalars_df, _ = scalars_dfs
+    left_columns = ["float64_col", "int64_too"]
+    right_columns = ["int64_col", "bool_col"]
+
+    left = scalars_df[left_columns]
+    right = scalars_df[right_columns]
+
+    with pytest.raises(
+        ValueError,
+        match="No common columns to perform merge on.",
+    ):
+        bpd.merge(left, right, "inner")
+
+
+def test_merge_raises_error_when_left_right_on_set(scalars_dfs):
+    scalars_df, _ = scalars_dfs
+    left_columns = ["int64_col", "int64_too"]
+    right_columns = ["int64_col", "bool_col"]
+
+    left = scalars_df[left_columns]
+    right = scalars_df[right_columns]
+
+    with pytest.raises(
+        ValueError,
+        match=re.escape("Can not pass both `on` and `left_on` + `right_on` params."),
+    ):
+        bpd.merge(
+            left,
+            right,
+            "inner",
+            left_on="int64_too",
+            right_on="int64_col",
+            on="int64_col",
+        )
+
+
 def _convert_pandas_category(pd_s: pd.Series):
     """
     Transforms a pandas Series with Categorical dtype into a bigframes-compatible