From af742a0654d48a87c577e92cc94d8f2e596c7b32 Mon Sep 17 00:00:00 2001
From: Ti Chi Robot <ti-community-prow-bot@tidb.io>
Date: Thu, 31 Oct 2024 17:46:28 +0800
Subject: [PATCH] Fix the wrong result bug when casting string as datetime with
 time zone or illegal chars (#9255) (#9351)

close pingcap/tiflash#8754

Fix the wrong result bug when casting string as datetime with time zone or illegal chars
- func `getFracIndex` ignore time zone.
- func `parseFrac` ignore illegal suffix after microsecond format.

Signed-off-by: Zhigao Tong <tongzhigao@pingcap.com>

Co-authored-by: TONG,Zhigao <tongzhigao@pingcap.com>
Co-authored-by: Zhigao Tong <tongzhigao@pingcap.com>
Co-authored-by: TONG, Zhigao <tongzhigao@pingcap.com>
---
 dbms/src/Common/MyTime.cpp             | 28 ++++++++++++++++++--------
 dbms/src/Common/tests/gtest_mytime.cpp |  5 +++++
 2 files changed, 25 insertions(+), 8 deletions(-)

diff --git a/dbms/src/Common/MyTime.cpp b/dbms/src/Common/MyTime.cpp
index 4f96592d226..a1e54a2004e 100644
--- a/dbms/src/Common/MyTime.cpp
+++ b/dbms/src/Common/MyTime.cpp
@@ -262,7 +262,7 @@ int getFracIndex(const String & format)
     int idx = -1;
     for (int i = end; i >= 0; i--)
     {
-        if (isPunctuation(format[i]))
+        if (format[i] != '+' && format[i] != '-' && isPunctuation(format[i]))
         {
             if (format[i] == '.')
             {
@@ -745,7 +745,7 @@ inline bool numberToDateTime(Int64 number, MyDateTime & result, bool allowZeroDa
 }
 
 // returns frac, overflow, matched. eg., "999" fsp=2 will overflow.
-std::tuple<UInt32, bool, bool> parseFrac(const String & str, int8_t fsp)
+std::tuple<UInt32, bool, bool> parseFrac(const std::string_view str, int8_t fsp)
 {
     if (str.empty())
     {
@@ -761,15 +761,27 @@ std::tuple<UInt32, bool, bool> parseFrac(const String & str, int8_t fsp)
     }
     try
     {
-        int len = str.length();
-        if (fsp >= len)
+        int end_pos = static_cast<size_t>(fsp) >= str.size() ? str.size() : (fsp + 1);
+        UInt32 tmp = 0;
+        int size = 0;
+        for (int i = 0; i < end_pos; ++i)
         {
-            UInt32 tmp = std::stoul(str);
-            return {tmp * std::pow(10, 6 - len), false, true};
+            if (auto c = str[i]; c >= '0' && c <= '9')
+            {
+                tmp = tmp * 10 + c - '0';
+                size++;
+            }
+            else
+            {
+                break;
+            }
+        }
+
+        if (fsp >= size)
+        {
+            return {tmp * std::pow(10, 6 - size), false, true};
         }
 
-        // Round when fsp < string length.
-        UInt32 tmp = std::stoul(str.substr(0, fsp + 1));
         tmp = (tmp + 5) / 10;
         if (tmp >= std::pow(10, fsp))
         {
diff --git a/dbms/src/Common/tests/gtest_mytime.cpp b/dbms/src/Common/tests/gtest_mytime.cpp
index c65a7dc05a0..5c076739128 100644
--- a/dbms/src/Common/tests/gtest_mytime.cpp
+++ b/dbms/src/Common/tests/gtest_mytime.cpp
@@ -125,6 +125,8 @@ try
         {"2020-10-10 10.10", "2020-10-10 10:10:00.000000"},
         {"2018.01.01", "2018-01-01 00:00:00.000000"},
         {"2020--12-10 11:11:11..123456", "2020-12-10 11:11:11.123456"},
+        {"2020-01-01 12:00:00.1234xxxx -0600 PST", "2020-01-01 12:00:00.123400"},
+        {"2020-01-01 12:00:00.123456-05:00", "2020-01-01 17:00:00.123456"}, // tidb/issues/49555
     };
     DataTypeMyDateTime type_with_fraction(6);
     for (auto & [str, expected] : cases_with_fsp)
@@ -184,6 +186,9 @@ try
     }
     DataTypeMyDateTime tp(2);
     checkParseMyDateTime("2010-12-31 23:59:59.99999", "2011-01-01 00:00:00.00", tp);
+    checkParseMyDateTime("2010-12-31 23:59:59.99xxxxx -0600 PST", "2010-12-31 23:59:59.99", tp);
+    checkParseMyDateTime("2020-01-01 12:00:00.123456 +0600 PST", "2020-01-01 12:00:00.12", tp);
+    checkParseMyDateTime("2020-01-01 12:00:00.123456 -0600 PST", "2020-01-01 12:00:00.12", tp);
 }
 catch (Exception & e)
 {