EHN: Add errors option in pandas.DataFrame.to_csv (#27750)

errors : str, default 'strict'
Behavior when the input string can’t be converted according to
the encoding’s rules (strict, ignore, replace, etc.)
See: https://docs.python.org/3/library/codecs.html#codec-base-classes

Author: shigemk2

doc/source/whatsnew/v1.0.0.rst

@@ -21,7 +21,7 @@ including other versions of pandas.
 Enhancements
 ~~~~~~~~~~~~
 
--
+- :meth:`Dataframe.to_csv` Add `encoding_errors` option (:issue:`27750`).
 -
 
 .. _whatsnew_1000.enhancements.other:

pandas/core/generic.py

@@ -3069,6 +3069,7 @@ def to_csv(
         doublequote=True,
         escapechar=None,
         decimal=".",
+        encoding_errors="strict",
     ):
         r"""
         Write object to a comma-separated values (csv) file.
@@ -3151,6 +3152,10 @@ def to_csv(
         decimal : str, default '.'
             Character recognized as decimal separator. E.g. use ',' for
             European data.
+        encoding_errors : str, default 'strict'
+            Behavior when the input string can’t be converted according to
+            the encoding’s rules (strict, ignore, replace, etc.)
+            See: https://docs.python.org/3/library/codecs.html#codec-base-classes
 
         Returns
         -------
@@ -3197,6 +3202,7 @@ def to_csv(
             doublequote=doublequote,
             escapechar=escapechar,
             decimal=decimal,
+            encoding_errors=encoding_errors,
         )
         formatter.save()
 

pandas/io/common.py

@@ -310,7 +310,13 @@ def _infer_compression(
 
 
 def _get_handle(
-    path_or_buf, mode, encoding=None, compression=None, memory_map=False, is_text=True
+    path_or_buf,
+    mode,
+    encoding=None,
+    compression=None,
+    memory_map=False,
+    is_text=True,
+    encoding_errors="strict",
 ):
     """
     Get file handle for given path/buffer and mode.
@@ -331,6 +337,10 @@ def _get_handle(
     is_text : boolean, default True
         whether file/buffer is in text format (csv, json, etc.), or in binary
         mode (pickle, etc.)
+    encoding_errors : str, default 'strict'
+        Behavior when the input string can’t be converted according to
+        the encoding’s rules (strict, ignore, replace, etc.)
+        See: https://docs.python.org/3/library/codecs.html#codec-base-classes
 
     Returns
     -------
@@ -407,10 +417,10 @@ def _get_handle(
     elif is_path:
         if encoding:
             # Encoding
-            f = open(path_or_buf, mode, encoding=encoding, newline="")
+            f = open(path_or_buf, mode, errors=encoding_errors, encoding=encoding, newline="")
         elif is_text:
             # No explicit encoding
-            f = open(path_or_buf, mode, errors="replace", newline="")
+            f = open(path_or_buf, mode, errors=encoding_errors, newline="")
         else:
             # Binary mode
             f = open(path_or_buf, mode)

pandas/io/formats/csvs.py

@@ -51,6 +51,7 @@ def __init__(
         doublequote=True,
         escapechar=None,
         decimal=".",
+        encoding_errors="strict",
     ):
 
         self.obj = obj
@@ -93,6 +94,8 @@ def __init__(
 
         self.has_mi_columns = isinstance(obj.columns, ABCMultiIndex)
 
+        self.encoding_errors = encoding_errors
+
         # validate mi options
         if self.has_mi_columns:
             if cols is not None:
@@ -179,6 +182,7 @@ def save(self):
                 self.mode,
                 encoding=self.encoding,
                 compression=self.compression,
+                encoding_errors=self.encoding_errors,
             )
             close = True