Adding overwrite or ignore behavior for row conflicts.

pandas/core/generic.py

@@ -2773,6 +2773,7 @@ def to_sql(
         con,
         schema=None,
         if_exists: str = "fail",
+        on_row_conflict: str = "fail",
         index: bool_t = True,
         index_label=None,
         chunksize=None,
@@ -2806,6 +2807,15 @@ def to_sql(
             * replace: Drop the table before inserting new values.
             * append: Insert new values to the existing table.
 
+        on_row_conflict : {'fail', 'replace', 'ignore'}, default 'fail'
+            How to behave if a row already exists.
+
+            * fail: Raise ValueError.
+            * replace: Overwrite the row with the incoming data.
+            * ignore: Ignore new data and keep existing data.
+
+            .. versionadded:: 1.5.0
+
         index : bool, default True
             Write DataFrame index as a column. Uses `index_label` as the column
             name in the table.
@@ -2850,6 +2860,9 @@ def to_sql(
         ValueError
             When the table already exists and `if_exists` is 'fail' (the
             default).
+        ValueError
+            When the row already exists and `on_row_conflict` is 'fail' (the
+            default).
 
         See Also
         --------
@@ -2942,6 +2955,7 @@ def to_sql(
             con,
             schema=schema,
             if_exists=if_exists,
+            on_row_conflict=on_row_conflict,
             index=index,
             index_label=index_label,
             chunksize=chunksize,