Misleading error for pd.read_msgpack (#27201)

Author: Kane

doc/source/whatsnew/v0.25.0.rst

@@ -1094,6 +1094,7 @@ I/O
 - Bug while selecting from :class:`HDFStore` with ``where=''`` specified (:issue:`26610`).
 - Fixed bug in :func:`DataFrame.to_excel()` where custom objects (i.e. `PeriodIndex`) inside merged cells were not being converted into types safe for the Excel writer (:issue:`27006`)
 - Bug in :meth:`read_hdf` where reading a timezone aware :class:`DatetimeIndex` would raise a ``TypeError`` (:issue:`11926`)
+- Bug in :meth:`to_msgpack` and :meth:`read_msgpack` which would raise a ``ValueError`` rather than a ``FileNotFoundError`` for an invalid path (:issue:`27160`)
 
 Plotting
 ^^^^^^^^

pandas/core/generic.py

@@ -2560,17 +2560,17 @@ def to_msgpack(self, path_or_buf=None, encoding="utf-8", **kwargs):
         Parameters
         ----------
         path : string File path, buffer-like, or None
-            if None, return generated string
+            if None, return generated bytes
         append : bool whether to append to an existing msgpack
             (default is False)
         compress : type of compressor (zlib or blosc), default to None (no
             compression)
 
         Returns
         -------
-        None or str
+        None or bytes
             If path_or_buf is None, returns the resulting msgpack format as a
-            string. Otherwise returns None.
+            byte string. Otherwise returns None.
         """
 
         from pandas.io import packers

pandas/io/packers.py

@@ -108,7 +108,7 @@ def to_msgpack(path_or_buf, *args, **kwargs):
     Parameters
     ----------
     path_or_buf : string File path, buffer-like, or None
-                  if None, return generated string
+                  if None, return generated bytes
     args : an object or objects to serialize
     encoding : encoding for unicode objects
     append : boolean whether to append to an existing msgpack
@@ -139,8 +139,12 @@ def writer(fh):
 
     path_or_buf = _stringify_path(path_or_buf)
     if isinstance(path_or_buf, str):
-        with open(path_or_buf, mode) as fh:
-            writer(fh)
+        try:
+            with open(path_or_buf, mode) as fh:
+                writer(fh)
+        except FileNotFoundError:
+            msg = "File b'{}' does not exist".format(path_or_buf)
+            raise FileNotFoundError(msg)
     elif path_or_buf is None:
         buf = BytesIO()
         writer(buf)
@@ -204,13 +208,11 @@ def read(fh):
     # see if we have an actual file
     if isinstance(path_or_buf, str):
         try:
-            exists = os.path.exists(path_or_buf)
-        except (TypeError, ValueError):
-            exists = False
-
-        if exists:
             with open(path_or_buf, "rb") as fh:
                 return read(fh)
+        except FileNotFoundError:
+            msg = "File b'{}' does not exist".format(path_or_buf)
+            raise FileNotFoundError(msg)
 
     if isinstance(path_or_buf, bytes):
         # treat as a binary-like

pandas/tests/io/test_common.py

@@ -142,7 +142,7 @@ def test_iterator(self):
             (pd.read_stata, "os", FileNotFoundError, "dta"),
             (pd.read_sas, "os", FileNotFoundError, "sas7bdat"),
             (pd.read_json, "os", ValueError, "json"),
-            (pd.read_msgpack, "os", ValueError, "mp"),
+            (pd.read_msgpack, "os", FileNotFoundError, "mp"),
             (pd.read_pickle, "os", FileNotFoundError, "pickle"),
         ],
     )
@@ -177,7 +177,7 @@ def test_read_non_existant(self, reader, module, error_class, fn_ext):
             (pd.read_stata, "os", FileNotFoundError, "dta"),
             (pd.read_sas, "os", FileNotFoundError, "sas7bdat"),
             (pd.read_json, "os", ValueError, "json"),
-            (pd.read_msgpack, "os", ValueError, "mp"),
+            (pd.read_msgpack, "os", FileNotFoundError, "mp"),
             (pd.read_pickle, "os", FileNotFoundError, "pickle"),
         ],
     )

pandas/tests/io/test_packers.py

@@ -165,12 +165,15 @@ def __init__(self):
                 self.read = 0
 
         msg = "Invalid file path or buffer object type: <class '{}'>"
+        invalid_path = os.path.join("nonexistent_dir", "df.msgpack")
         with pytest.raises(ValueError, match=msg.format("NoneType")):
             read_msgpack(path_or_buf=None)
         with pytest.raises(ValueError, match=msg.format("dict")):
             read_msgpack(path_or_buf={})
         with pytest.raises(ValueError, match=msg.format(r".*\.A")):
             read_msgpack(path_or_buf=A())
+        with pytest.raises(FileNotFoundError, match="does not exist"):
+            read_msgpack(path_or_buf=invalid_path)
 
 
 @pytest.mark.filterwarnings("ignore:.*msgpack:FutureWarning")