Fix issue pandas-dev#36271 - pd.read_json() fails for strings that lo…

…ok similar to fsspec_url (pandas-dev#44619)

doc/source/whatsnew/v1.4.0.rst

@@ -785,6 +785,7 @@ I/O
 - Bug in :func:`read_csv` not setting name of :class:`MultiIndex` columns correctly when ``index_col`` is not the first column (:issue:`38549`)
 - Bug in :func:`read_csv` silently ignoring errors when failing to create a memory-mapped file (:issue:`44766`)
 - Bug in :func:`read_csv` when passing a ``tempfile.SpooledTemporaryFile`` opened in binary mode (:issue:`44748`)
+- Bug in :func:`read_json` raising ``ValueError`` when attempting to parse json strings containing "://" (:issue:`36271`)
 -
 
 Period

pandas/io/common.py

@@ -18,6 +18,7 @@
 import mmap
 import os
 from pathlib import Path
+import re
 from typing import (
     IO,
     Any,
@@ -59,6 +60,7 @@
 
 _VALID_URLS = set(uses_relative + uses_netloc + uses_params)
 _VALID_URLS.discard("")
+_RFC_3986_PATTERN = re.compile(r"^[A-Za-z][A-Za-z0-9+\-+.]*://")
 
 BaseBufferT = TypeVar("BaseBufferT", bound=BaseBuffer)
 
@@ -244,7 +246,7 @@ def is_fsspec_url(url: FilePath | BaseBuffer) -> bool:
     """
     return (
         isinstance(url, str)
-        and "://" in url
+        and bool(_RFC_3986_PATTERN.match(url))
         and not url.startswith(("http://", "https://"))
     )
 

pandas/tests/io/json/test_pandas.py

@@ -1527,6 +1527,21 @@ def test_read_timezone_information(self):
         expected = Series([88], index=DatetimeIndex(["2019-01-01 11:00:00"], tz="UTC"))
         tm.assert_series_equal(result, expected)
 
+    @pytest.mark.parametrize(
+        "url",
+        [
+            "s3://example-fsspec/",
+            "gcs://another-fsspec/file.json",
+            "https://example-site.com/data",
+            "some-protocol://data.txt",
+        ],
+    )
+    def test_read_json_with_url_value(self, url):
+        # GH 36271
+        result = read_json(f'{{"url":{{"0":"{url}"}}}}')
+        expected = DataFrame({"url": [url]})
+        tm.assert_frame_equal(result, expected)
+
     @pytest.mark.parametrize(
         "date_format,key", [("epoch", 86400000), ("iso", "P1DT0H0M0S")]
     )

pandas/tests/io/test_common.py

@@ -494,6 +494,11 @@ def test_is_fsspec_url():
     assert not icom.is_fsspec_url("random:pandas/somethingelse.com")
     assert not icom.is_fsspec_url("/local/path")
     assert not icom.is_fsspec_url("relative/local/path")
+    # fsspec URL in string should not be recognized
+    assert not icom.is_fsspec_url("this is not fsspec://url")
+    assert not icom.is_fsspec_url("{'url': 'gs://pandas/somethingelse.com'}")
+    # accept everything that conforms to RFC 3986 schema
+    assert icom.is_fsspec_url("RFC-3986+compliant.spec://something")
 
 
 @pytest.mark.parametrize("encoding", [None, "utf-8"])