From 21eb6f036ef3eac2d9f2468ec8f96ac5693af878 Mon Sep 17 00:00:00 2001
From: Rebecca Schlussel <rschlussel@fb.com>
Date: Thu, 8 Aug 2024 11:51:04 -0400
Subject: [PATCH] Fix like matching with newlines and no wildcard

LIKE expressions were not looking past newlines when there was no wild
card. That means that certain expressions were incorrectly returning
matches when after the newline the input did not match. This change
fixes that behavior. For example:

SELECT 'foo\nbar' LIKE 'foo'.

Previously that query would return "true". Now it will return
false.
---
 .../src/main/sphinx/functions/comparison.rst  |  5 ++++-
 .../facebook/presto/type/LikeFunctions.java   |  5 ++++-
 .../presto/sql/TestLikeFunctions.java         | 20 +++++++++++++++++++
 3 files changed, 28 insertions(+), 2 deletions(-)

diff --git a/presto-docs/src/main/sphinx/functions/comparison.rst b/presto-docs/src/main/sphinx/functions/comparison.rst
index 0c32ee6eb089c..62da5d7936ba8 100644
--- a/presto-docs/src/main/sphinx/functions/comparison.rst
+++ b/presto-docs/src/main/sphinx/functions/comparison.rst
@@ -158,7 +158,7 @@ LIKE
 ----
 The LIKE operator is used to match a specified character pattern in a string. Patterns can contain
 regular characters as well as wildcards. Wildcard characters can be escaped using the single character
-specified for the ESCAPE parameter. Matching is case sensitive.
+specified for the ESCAPE parameter. Matching is case sensitive, and the pattern must match the whole string.
 
 Syntax:
 
@@ -199,6 +199,9 @@ Examples::
     WHERE name LIKE '%#%%' ESCAPE '#'
     --returns 'a%c' and  '%cd'
 
+   SELECT 'ab' || chr(10) || 'c' LIKE 'ab'  --chr(10) is a newline character
+    --returns 'false'
+
 
 Row comparison: IN
 ------------------
diff --git a/presto-main/src/main/java/com/facebook/presto/type/LikeFunctions.java b/presto-main/src/main/java/com/facebook/presto/type/LikeFunctions.java
index 84c01c5b7327d..1673d2f9a8964 100644
--- a/presto-main/src/main/java/com/facebook/presto/type/LikeFunctions.java
+++ b/presto-main/src/main/java/com/facebook/presto/type/LikeFunctions.java
@@ -212,7 +212,10 @@ private static Regex likePattern(String patternString, char escapeChar, boolean
         regex.append('$');
 
         byte[] bytes = regex.toString().getBytes(UTF_8);
-        return new Regex(bytes, 0, bytes.length, Option.MULTILINE, NonStrictUTF8Encoding.INSTANCE, SYNTAX);
+        // Option.MULTILINE specifies that wildcard characters (. and *) should match newlines
+        // Option.SINGLELINE specifies that anchors (^ and $) should match the beginning and end of
+        // input rather than the beginning and end of the line
+        return new Regex(bytes, 0, bytes.length, Option.MULTILINE | Option.SINGLELINE, NonStrictUTF8Encoding.INSTANCE, SYNTAX);
     }
 
     @SuppressWarnings("NumericCastThatLosesPrecision")
diff --git a/presto-main/src/test/java/com/facebook/presto/sql/TestLikeFunctions.java b/presto-main/src/test/java/com/facebook/presto/sql/TestLikeFunctions.java
index 496cd578519bf..b0d75d23a12a1 100644
--- a/presto-main/src/test/java/com/facebook/presto/sql/TestLikeFunctions.java
+++ b/presto-main/src/test/java/com/facebook/presto/sql/TestLikeFunctions.java
@@ -96,12 +96,32 @@ public void testLikeNewlineBeforeMatch()
         assertTrue(likeVarchar(utf8Slice("foo\nbar"), regex));
     }
 
+    @Test
+    public void testLikeNewlineNoWildcard()
+    {
+        Regex regex = likePattern(utf8Slice("foo\nbar"));
+        assertTrue(likeVarchar(utf8Slice("foo\nbar"), regex));
+    }
+
+    @Test
+    public void testLikeNoMatchAfterNewline()
+    {
+        Regex regex = likePattern(utf8Slice("foo"));
+        assertFalse(likeVarchar(utf8Slice("foo\nbar"), regex));
+    }
+
     @Test
     public void testLikeNewlineInMatch()
     {
         Regex regex = likePattern(utf8Slice("f%b%"));
         assertTrue(likeVarchar(utf8Slice("foo\nbar"), regex));
     }
+    @Test
+    public void testLikeNewlineInSingleWildcardMatch()
+    {
+        Regex regex = likePattern(utf8Slice("foo_bar"));
+        assertTrue(likeVarchar(utf8Slice("foo\nbar"), regex));
+    }
 
     @Test(timeOut = 1000)
     public void testLikeUtf8Pattern()