From 2be8f83e9fa006232137a42bbbcfc2ba64afbfef Mon Sep 17 00:00:00 2001
From: "Miss Islington (bot)"
 <31488909+miss-islington@users.noreply.github.com>
Date: Tue, 4 Aug 2026 11:03:37 +0200
Subject: [PATCH 1/2] [3.10] gh-98820: Fix quadratic time in csv.Sniffer for
 quoted fields (GH-154867)

(cherry picked from commit b30c7fa9edd921a118f286e9f90f560777fa693b)
(cherry picked from commit 39dac551f9721da974ccfafaca6b632b3212fcb7)
---
 Lib/csv.py                                         | 14 +++++++++-----
 Lib/test/test_csv.py                               |  7 +++++++
 .../2026-07-29-11-20-00.gh-issue-98820.Qm7Hs4.rst  |  2 ++
 3 files changed, 18 insertions(+), 5 deletions(-)
 create mode 100644 Misc/NEWS.d/next/Library/2026-07-29-11-20-00.gh-issue-98820.Qm7Hs4.rst

diff --git a/Lib/csv.py b/Lib/csv.py
index 5866f3796c6a9cf..745367bddce11cb 100644
--- a/Lib/csv.py
+++ b/Lib/csv.py
@@ -215,12 +215,16 @@ def _guess_quote_and_delimiter(self, data, delimiters):
         this way.
         """
 
+        # The body of a quoted field ends at the first quote which is
+        # not doubled, as it does for a reader.  A lazy ".*?" scans to
+        # the end of the sample instead, from every start: quadratically.
+        body = r'(?:(?P=quote){2}|(?!(?P=quote)).)*+'
         matches = []
-        for restr in (r'(?P<delim>[^\w\n"\'])(?P<space> ?)(?P<quote>["\']).*?(?P=quote)(?P=delim)', # ,".*?",
-                      r'(?:^|\n)(?P<quote>["\']).*?(?P=quote)(?P<delim>[^\w\n"\'])(?P<space> ?)',   #  ".*?",
-                      r'(?P<delim>[^\w\n"\'])(?P<space> ?)(?P<quote>["\']).*?(?P=quote)(?:$|\n)',   # ,".*?"
-                      r'(?:^|\n)(?P<quote>["\']).*?(?P=quote)(?:$|\n)'):                            #  ".*?" (no delim, no space)
-            regexp = re.compile(restr, re.DOTALL | re.MULTILINE)
+        for restr in (r'(?P<delim>[^\w\n"\'])(?P<space> ?)(?P<quote>["\'])%s(?P=quote)(?P=delim)',   # ,"...",
+                      r'(?:^|\n)(?P<quote>["\'])%s(?P=quote)(?P<delim>[^\w\n"\'])(?P<space> ?)',     #  "...",
+                      r'(?P<delim>[^\w\n"\'])(?P<space> ?)(?P<quote>["\'])%s(?P=quote)(?:$|\n)',  # ,"..."
+                      r'(?:^|\n)(?P<quote>["\'])%s(?P=quote)(?:$|\n)'):                           #  "..." (no delim, no space)
+            regexp = re.compile(restr % body, re.DOTALL | re.MULTILINE)
             matches = regexp.findall(data)
             if matches:
                 break
diff --git a/Lib/test/test_csv.py b/Lib/test/test_csv.py
index 6ff5bc3d644e69d..f83929b5626939a 100644
--- a/Lib/test/test_csv.py
+++ b/Lib/test/test_csv.py
@@ -1232,6 +1232,13 @@ def test_sniff_space_delimiter(self):
         self.assertEqual(dialect.delimiter, ' ')
         self.assertIs(dialect.doublequote, False)
 
+    def test_sniff_quoted_single_column(self):
+        # gh-98820: this sample used to take minutes.
+        sniffer = csv.Sniffer()
+        sample = '"abcdefghijklmnopqrstuvwxyz"\n' * 10000
+        with self.assertRaisesRegex(csv.Error, "Could not determine delimiter"):
+            sniffer.sniff(sample, delimiters=',:|\t')
+
 
 class NUL:
     def write(s, *args):
diff --git a/Misc/NEWS.d/next/Library/2026-07-29-11-20-00.gh-issue-98820.Qm7Hs4.rst b/Misc/NEWS.d/next/Library/2026-07-29-11-20-00.gh-issue-98820.Qm7Hs4.rst
new file mode 100644
index 000000000000000..aa9ae8d937004fe
--- /dev/null
+++ b/Misc/NEWS.d/next/Library/2026-07-29-11-20-00.gh-issue-98820.Qm7Hs4.rst
@@ -0,0 +1,2 @@
+Fix quadratic time in :meth:`csv.Sniffer.sniff` for a sample which contains
+quoted fields, in particular for a single column of quoted fields.

From 413b9d6371f2f3da1e1127cde8979ab42f3fbb65 Mon Sep 17 00:00:00 2001
From: Serhiy Storchaka <storchaka@gmail.com>
Date: Tue, 11 Aug 2026 09:40:05 +0300
Subject: [PATCH 2/2] Match the body of a quoted field without possessive
 quantifiers

They were added in 3.11.  An unrolled loop is unambiguous at every
position, so it does not backtrack either.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---
 Lib/csv.py | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/Lib/csv.py b/Lib/csv.py
index 745367bddce11cb..4e95bf9eb743265 100644
--- a/Lib/csv.py
+++ b/Lib/csv.py
@@ -218,7 +218,9 @@ def _guess_quote_and_delimiter(self, data, delimiters):
         # The body of a quoted field ends at the first quote which is
         # not doubled, as it does for a reader.  A lazy ".*?" scans to
         # the end of the sample instead, from every start: quadratically.
-        body = r'(?:(?P=quote){2}|(?!(?P=quote)).)*+'
+        # As an unrolled loop it is unambiguous, so it does not backtrack.
+        other = r'(?:(?!(?P=quote)).)*'
+        body = r'%s(?:(?P=quote){2}%s)*' % (other, other)
         matches = []
         for restr in (r'(?P<delim>[^\w\n"\'])(?P<space> ?)(?P<quote>["\'])%s(?P=quote)(?P=delim)',   # ,"...",
                       r'(?:^|\n)(?P<quote>["\'])%s(?P=quote)(?P<delim>[^\w\n"\'])(?P<space> ?)',     #  "...",