BUG: read csv not breaking lines for warn messages (pandas-dev#45677)

Author: phofl

doc/source/whatsnew/v1.5.0.rst

@@ -286,6 +286,7 @@ I/O
 ^^^
 - Bug in :meth:`DataFrame.to_stata` where no error is raised if the :class:`DataFrame` contains ``-np.inf`` (:issue:`45350`)
 - Bug in :meth:`DataFrame.info` where a new line at the end of the output is omitted when called on an empty :class:`DataFrame` (:issue:`45494`)
+- Bug in :func:`read_csv` not recognizing line break for ``on_bad_lines="warn"`` for ``engine="c"`` (:issue:`41710`)
 - Bug in :func:`read_parquet` when ``engine="pyarrow"`` which caused partial write to disk when column of unsupported datatype was passed (:issue:`44914`)
 -
 

pandas/_libs/parsers.pyx

@@ -1,5 +1,6 @@
 # Copyright (c) 2012, Lambda Foundry, Inc.
 # See LICENSE for the license
+from base64 import decode
 from csv import (
     QUOTE_MINIMAL,
     QUOTE_NONE,
@@ -839,7 +840,9 @@ cdef class TextReader:
             status = tokenize_nrows(self.parser, nrows, self.encoding_errors)
 
         if self.parser.warn_msg != NULL:
-            print(self.parser.warn_msg, file=sys.stderr)
+            print(PyUnicode_DecodeUTF8(
+                self.parser.warn_msg, strlen(self.parser.warn_msg),
+                self.encoding_errors), file=sys.stderr)
             free(self.parser.warn_msg)
             self.parser.warn_msg = NULL
 
@@ -868,7 +871,9 @@ cdef class TextReader:
                 status = tokenize_all_rows(self.parser, self.encoding_errors)
 
             if self.parser.warn_msg != NULL:
-                print(self.parser.warn_msg, file=sys.stderr)
+                print(PyUnicode_DecodeUTF8(
+                    self.parser.warn_msg, strlen(self.parser.warn_msg),
+                    self.encoding_errors), file=sys.stderr)
                 free(self.parser.warn_msg)
                 self.parser.warn_msg = NULL
 

pandas/tests/io/parser/common/test_read_errors.py

@@ -278,3 +278,30 @@ def test_conflict_on_bad_line(all_parsers, error_bad_lines, warn_bad_lines):
         "Please only set on_bad_lines.",
     ):
         parser.read_csv(StringIO(data), on_bad_lines="error", **kwds)
+
+
+def test_on_bad_lines_warn_correct_formatting(all_parsers, capsys):
+    # see gh-15925
+    parser = all_parsers
+    data = """1,2
+a,b
+a,b,c
+a,b,d
+a,b
+"""
+    expected = DataFrame({"1": "a", "2": ["b"] * 2})
+
+    result = parser.read_csv(StringIO(data), on_bad_lines="warn")
+    tm.assert_frame_equal(result, expected)
+
+    captured = capsys.readouterr()
+    if parser.engine == "c":
+        warn = """Skipping line 3: expected 2 fields, saw 3
+Skipping line 4: expected 2 fields, saw 3
+
+"""
+    else:
+        warn = """Skipping line 3: Expected 2 fields in line 3, saw 3
+Skipping line 4: Expected 2 fields in line 4, saw 3
+"""
+    assert captured.err == warn