DEPR: error_bad_lines and warn_bad_lines for read_csv

doc/source/user_guide/io.rst

@@ -344,16 +344,33 @@ dialect : str or :class:`python:csv.Dialect` instance, default ``None``
 Error handling
 ++++++++++++++
 
-error_bad_lines : boolean, default ``True``
+error_bad_lines : boolean, default ``None``
   Lines with too many fields (e.g. a csv line with too many commas) will by
   default cause an exception to be raised, and no ``DataFrame`` will be
   returned. If ``False``, then these "bad lines" will dropped from the
   ``DataFrame`` that is returned. See :ref:`bad lines <io.bad_lines>`
   below.
-warn_bad_lines : boolean, default ``True``
+
+  .. deprecated:: 1.3
+     The ``on_bad_lines`` parameter should be used instead to specify behavior upon
+     encountering a bad line instead.
+warn_bad_lines : boolean, default ``None``
   If error_bad_lines is ``False``, and warn_bad_lines is ``True``, a warning for
   each "bad line" will be output.
 
+  .. deprecated:: 1.3
+     The ``on_bad_lines`` parameter should be used instead to specify behavior upon
+     encountering a bad line instead.
+on_bad_lines : {{'error', 'warn', 'skip'}}, default 'error'
+    Specifies what to do upon encountering a bad line (a line with too many fields).
+    Allowed values are :
+
+        - 'error', raise an ParserError when a bad line is encountered.
+        - 'warn', print a warning when a bad line is encountered and skip that line.
+        - 'skip', skip bad lines without raising or warning when they are encountered.
+
+    .. versionadded:: 1.3
+
 .. _io.dtypes:
 
 Specifying column data types
@@ -1245,7 +1262,7 @@ You can elect to skip bad lines:
 
 .. code-block:: ipython
 
-    In [29]: pd.read_csv(StringIO(data), error_bad_lines=False)
+    In [29]: pd.read_csv(StringIO(data), on_bad_lines="warn")
     Skipping line 3: expected 3 fields, saw 4
 
     Out[29]:

doc/source/whatsnew/v1.3.0.rst

@@ -669,6 +669,7 @@ Deprecations
 - Deprecated casting ``datetime.date`` objects to ``datetime64`` when used as ``fill_value`` in :meth:`DataFrame.unstack`, :meth:`DataFrame.shift`, :meth:`Series.shift`, and :meth:`DataFrame.reindex`, pass ``pd.Timestamp(dateobj)`` instead (:issue:`39767`)
 - Deprecated :meth:`.Styler.set_na_rep` and :meth:`.Styler.set_precision` in favour of :meth:`.Styler.format` with ``na_rep`` and ``precision`` as existing and new input arguments respectively (:issue:`40134`, :issue:`40425`)
 - Deprecated allowing partial failure in :meth:`Series.transform` and :meth:`DataFrame.transform` when ``func`` is list-like or dict-like and raises anything but ``TypeError``; ``func`` raising anything but a ``TypeError`` will raise in a future version (:issue:`40211`)
+- Deprecated arguments ``error_bad_lines`` and ``warn_bad_lines`` in :meth:``read_csv`` and :meth:``read_table`` in favor of argument ``on_bad_lines`` (:issue:`15122`)
 - Deprecated support for ``np.ma.mrecords.MaskedRecords`` in the :class:`DataFrame` constructor, pass ``{name: data[name] for name in data.dtype.names}`` instead (:issue:`40363`)
 - Deprecated using :func:`merge` or :func:`join` on a different number of levels (:issue:`34862`)
 - Deprecated the use of ``**kwargs`` in :class:`.ExcelWriter`; use the keyword argument ``engine_kwargs`` instead (:issue:`40430`)

pandas/_libs/parsers.pyx

@@ -146,6 +146,11 @@ cdef extern from "parser/tokenizer.h":
 
     enum: ERROR_OVERFLOW
 
+    ctypedef enum BadLineHandleMethod:
+        ERROR,
+        WARN,
+        SKIP
+
     ctypedef void* (*io_callback)(void *src, size_t nbytes, size_t *bytes_read,
                                   int *status, const char *encoding_errors)
     ctypedef int (*io_cleanup)(void *src)
@@ -198,8 +203,7 @@ cdef extern from "parser/tokenizer.h":
         int usecols
 
         int expected_fields
-        int error_bad_lines
-        int warn_bad_lines
+        BadLineHandleMethod on_bad_lines
 
         # floating point options
         char decimal
@@ -351,8 +355,7 @@ cdef class TextReader:
                   thousands=None,       # bytes | str
                   dtype=None,
                   usecols=None,
-                  bint error_bad_lines=True,
-                  bint warn_bad_lines=True,
+                  on_bad_lines = ERROR,
                   bint na_filter=True,
                   na_values=None,
                   na_fvalues=None,
@@ -435,9 +438,7 @@ cdef class TextReader:
                 raise ValueError('Only length-1 comment characters supported')
             self.parser.commentchar = ord(comment)
 
-        # error handling of bad lines
-        self.parser.error_bad_lines = int(error_bad_lines)
-        self.parser.warn_bad_lines = int(warn_bad_lines)
+        self.parser.on_bad_lines = on_bad_lines
 
         self.skiprows = skiprows
         if skiprows is not None:
@@ -454,8 +455,7 @@ cdef class TextReader:
 
         # XXX
         if skipfooter > 0:
-            self.parser.error_bad_lines = 0
-            self.parser.warn_bad_lines = 0
+            self.parser.on_bad_lines = SKIP
 
         self.delimiter = delimiter
 
@@ -570,9 +570,6 @@ cdef class TextReader:
             kh_destroy_str_starts(self.false_set)
             self.false_set = NULL
 
-    def set_error_bad_lines(self, int status) -> None:
-        self.parser.error_bad_lines = status
-
     def _set_quoting(self, quote_char: str | bytes | None, quoting: int):
         if not isinstance(quoting, int):
             raise TypeError('"quoting" must be an integer')

pandas/_libs/src/parser/tokenizer.c

@@ -93,8 +93,7 @@ void parser_set_default_options(parser_t *self) {
     self->allow_embedded_newline = 1;
 
     self->expected_fields = -1;
-    self->error_bad_lines = 0;
-    self->warn_bad_lines = 0;
+    self->on_bad_lines = ERROR;
 
     self->commentchar = '#';
     self->thousands = '\0';
@@ -457,7 +456,7 @@ static int end_line(parser_t *self) {
         self->line_fields[self->lines] = 0;
 
         // file_lines is now the actual file line number (starting at 1)
-        if (self->error_bad_lines) {
+        if (self->on_bad_lines == ERROR) {
             self->error_msg = malloc(bufsize);
             snprintf(self->error_msg, bufsize,
                     "Expected %d fields in line %" PRIu64 ", saw %" PRId64 "\n",
@@ -468,7 +467,7 @@ static int end_line(parser_t *self) {
             return -1;
         } else {
             // simply skip bad lines
-            if (self->warn_bad_lines) {
+            if (self->on_bad_lines == WARN) {
                 // pass up error message
                 msg = malloc(bufsize);
                 snprintf(msg, bufsize,

pandas/_libs/src/parser/tokenizer.h

@@ -84,6 +84,12 @@ typedef enum {
     QUOTE_NONE
 } QuoteStyle;
 
+typedef enum {
+    ERROR,
+    WARN,
+    SKIP
+} BadLineHandleMethod;
+
 typedef void *(*io_callback)(void *src, size_t nbytes, size_t *bytes_read,
                              int *status, const char *encoding_errors);
 typedef int (*io_cleanup)(void *src);
@@ -136,8 +142,7 @@ typedef struct parser_t {
     int usecols;  // Boolean: 1: usecols provided, 0: none provided
 
     int expected_fields;
-    int error_bad_lines;
-    int warn_bad_lines;
+    BadLineHandleMethod on_bad_lines;
 
     // floating point options
     char decimal;

pandas/io/parsers/base_parser.py

@@ -3,6 +3,7 @@
 from collections import defaultdict
 import csv
 import datetime
+from enum import Enum
 import itertools
 from typing import (
     Any,
@@ -108,10 +109,16 @@
     "infer_datetime_format": False,
     "skip_blank_lines": True,
     "encoding_errors": "strict",
+    "on_bad_lines": "error",
 }
 
 
 class ParserBase:
+    class BadLineHandleMethod(Enum):
+        ERROR = 0
+        WARN = 1
+        SKIP = 2
+
     _implicit_index: bool = False
     _first_chunk: bool
 
@@ -203,9 +210,13 @@ def __init__(self, kwds):
 
         self.handles: IOHandles | None = None
 
+        # Fallback to error to pass a sketchy test(test_override_set_noconvert_columns)
+        # Normally, this arg would get pre-processed earlier on
+        self.on_bad_lines = kwds.get("on_bad_lines", self.BadLineHandleMethod.ERROR)
+
     def _open_handles(self, src: FilePathOrBuffer, kwds: dict[str, Any]) -> None:
         """
-        Let the readers open IOHanldes after they are done with their potential raises.
+        Let the readers open IOHandles after they are done with their potential raises.
         """
         self.handles = get_handle(
             src,

pandas/io/parsers/c_parser_wrapper.py

@@ -50,7 +50,18 @@ def __init__(self, src: FilePathOrBuffer, **kwds):
         # open handles
         self._open_handles(src, kwds)
         assert self.handles is not None
-        for key in ("storage_options", "encoding", "memory_map", "compression"):
+
+        # Have to pass int, would break tests using TextReader directly otherwise :(
+        kwds["on_bad_lines"] = self.on_bad_lines.value
+
+        for key in (
+            "storage_options",
+            "encoding",
+            "memory_map",
+            "compression",
+            "error_bad_lines",
+            "warn_bad_lines",
+        ):
             kwds.pop(key, None)
 
         kwds["dtype"] = ensure_dtype_objs(kwds.get("dtype", None))
@@ -206,9 +217,6 @@ def _set_noconvert_columns(self):
         for col in noconvert_columns:
             self._reader.set_noconvert(col)
 
-    def set_error_bad_lines(self, status):
-        self._reader.set_error_bad_lines(int(status))
-
     def read(self, nrows=None):
         try:
             if self.low_memory:

pandas/io/parsers/python_parser.py

@@ -74,9 +74,6 @@ def __init__(self, f: Union[FilePathOrBuffer, list], **kwds):
         self.quoting = kwds["quoting"]
         self.skip_blank_lines = kwds["skip_blank_lines"]
 
-        self.warn_bad_lines = kwds["warn_bad_lines"]
-        self.error_bad_lines = kwds["error_bad_lines"]
-
         self.names_passed = kwds["names"] or None
 
         self.has_index_names = False
@@ -707,10 +704,11 @@ def _next_line(self):
 
     def _alert_malformed(self, msg, row_num):
         """
-        Alert a user about a malformed row.
+        Alert a user about a malformed row, depending on value of
+        `self.on_bad_lines` enum.
 
-        If `self.error_bad_lines` is True, the alert will be `ParserError`.
-        If `self.warn_bad_lines` is True, the alert will be printed out.
+        If `self.on_bad_lines` is ERROR, the alert will be `ParserError`.
+        If `self.on_bad_lines` is WARN, the alert will be printed out.
 
         Parameters
         ----------
@@ -719,9 +717,9 @@ def _alert_malformed(self, msg, row_num):
                   Because this row number is displayed, we 1-index,
                   even though we 0-index internally.
         """
-        if self.error_bad_lines:
+        if self.on_bad_lines == self.BadLineHandleMethod.ERROR:
             raise ParserError(msg)
-        elif self.warn_bad_lines:
+        elif self.on_bad_lines == self.BadLineHandleMethod.WARN:
             base = f"Skipping line {row_num}: "
             sys.stderr.write(base + msg + "\n")
 
@@ -742,7 +740,10 @@ def _next_iter_line(self, row_num):
             assert self.data is not None
             return next(self.data)
         except csv.Error as e:
-            if self.warn_bad_lines or self.error_bad_lines:
+            if (
+                self.on_bad_lines == self.BadLineHandleMethod.ERROR
+                or self.on_bad_lines == self.BadLineHandleMethod.WARN
+            ):
                 msg = str(e)
 
                 if "NULL byte" in msg or "line contains NUL" in msg:
@@ -947,11 +948,14 @@ def _rows_to_cols(self, content):
                 actual_len = len(l)
 
                 if actual_len > col_len:
-                    if self.error_bad_lines or self.warn_bad_lines:
+                    if (
+                        self.on_bad_lines == self.BadLineHandleMethod.ERROR
+                        or self.on_bad_lines == self.BadLineHandleMethod.WARN
+                    ):
                         row_num = self.pos - (content_len - i + footers)
                         bad_lines.append((row_num, actual_len))
 
-                        if self.error_bad_lines:
+                        if self.on_bad_lines == self.BadLineHandleMethod.ERROR:
                             break
                 else:
                     content.append(l)