BUG: read_csv with chained fsspec TAR file and compression="infer" (#60100)

Author: KevsterAmp

doc/source/whatsnew/v3.0.0.rst

@@ -784,6 +784,7 @@ Other
 - Bug in :meth:`Series.dt` methods in :class:`ArrowDtype` that were returning incorrect values. (:issue:`57355`)
 - Bug in :meth:`Series.rank` that doesn't preserve missing values for nullable integers when ``na_option='keep'``. (:issue:`56976`)
 - Bug in :meth:`Series.replace` and :meth:`DataFrame.replace` inconsistently replacing matching instances when ``regex=True`` and missing values are present. (:issue:`56599`)
+- Bug in :meth:`read_csv` where chained fsspec TAR file and ``compression="infer"`` fails with ``tarfile.ReadError`` (:issue:`60028`)
 - Bug in Dataframe Interchange Protocol implementation was returning incorrect results for data buffers' associated dtype, for string and datetime columns (:issue:`54781`)
 - Bug in ``Series.list`` methods not preserving the original :class:`Index`. (:issue:`58425`)
 

pandas/io/common.py

@@ -584,6 +584,9 @@ def infer_compression(
     # Infer compression
     if compression == "infer":
         # Convert all path types (e.g. pathlib.Path) to strings
+        if isinstance(filepath_or_buffer, str) and "::" in filepath_or_buffer:
+            # chained URLs contain ::
+            filepath_or_buffer = filepath_or_buffer.split("::")[0]
         filepath_or_buffer = stringify_path(filepath_or_buffer, convert_file_like=True)
         if not isinstance(filepath_or_buffer, str):
             # Cannot infer compression of a buffer, assume no compression

pandas/tests/io/data/tar/test-csv.tar

@@ -25,6 +25,7 @@
     WASM,
     is_platform_windows,
 )
+import pandas.util._test_decorators as td
 
 import pandas as pd
 import pandas._testing as tm
@@ -642,6 +643,19 @@ def close(self):
                 handles.created_handles.append(TestError())
 
 
+@td.skip_if_no("fsspec", min_version="2023.1.0")
+@pytest.mark.parametrize("compression", [None, "infer"])
+def test_read_csv_chained_url_no_error(compression):
+    # GH 60100
+    tar_file_path = "pandas/tests/io/data/tar/test-csv.tar"
+    chained_file_url = f"tar://test.csv::file://{tar_file_path}"
+
+    result = pd.read_csv(chained_file_url, compression=compression, sep=";")
+    expected = pd.DataFrame({"1": {0: 3}, "2": {0: 4}})
+
+    tm.assert_frame_equal(expected, result)
+
+
 @pytest.mark.parametrize(
     "reader",
     [