[GitHub] [arrow] ianmcook commented on a change in pull request #10356: ARROW-12715: [C++][Python] Add SQL LIKE match kernel

[GitBox <gi...@apache.org>]

ianmcook commented on a change in pull request #10356:
URL: https://github.com/apache/arrow/pull/10356#discussion_r635579350



##########
File path: cpp/src/arrow/compute/kernels/scalar_string.cc
##########
@@ -494,6 +494,93 @@ const FunctionDoc match_substring_regex_doc(
      "position.\n"
      "Null inputs emit null.  The pattern must be given in MatchSubstringOptions."),
     {"strings"}, "MatchSubstringOptions");
+
+// SQL LIKE match
+
+/// Convert a SQL-style LIKE pattern (using '%' and '_') into a regex pattern
+std::string MakeLikeRegex(const MatchSubstringOptions& options) {
+  // Allow . to match \n
+  std::string like_pattern = "(?s:^";
+  like_pattern.reserve(options.pattern.size() + 7);
+  bool escaped = false;
+  for (const char c : options.pattern) {
+    if (!escaped && c == '%') {
+      like_pattern.append(".*");
+    } else if (!escaped && c == '_') {
+      like_pattern.append(".");
+    } else if (!escaped && c == '\\') {
+      escaped = true;
+    } else {
+      switch (c) {
+        case '.':
+        case '?':
+        case '+':
+        case '*':
+        case '^':
+        case '$':
+        case '\\':
+        case '[':
+        case '{':
+        case '(':
+        case ')':
+        case '|': {

Review comment:
       I'm pretty sure we don't need to escape `]` and `}` but I noticed Impala does:
   https://github.com/apache/impala/blob/9c38568657d62b6f6d7b10aa1c721ba843374dd8/be/src/exprs/like-predicate.cc#L384-L397
   Can you think of any reason we would need to escape those?




-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org