fix: honor offset when occurence is 0 (#450)

GoogleCloudPlatform · Oct 1, 2024 · ba69b58 · ba69b58
1 parent 1b05b09
commit ba69b58
Show file tree

Hide file tree

Showing 2 changed files with 20 additions and 2 deletions.
diff --git a/udfs/community/cw_regexp_replace_generic.sqlx b/udfs/community/cw_regexp_replace_generic.sqlx
@@ -21,9 +21,9 @@ LANGUAGE js AS """
   if (haystack == null || regexp == null || replacement == null || offset == null || occurrence == null || mode == null) return null;
   replacement = replacement.replace('\\\\', '$');
   let regExp = new RegExp(regexp, mode);
-  if (occurrence == 0)
-      return haystack.replace(regExp, replacement);
   const start = offset - 1;
+  if (occurrence == 0)
+      return haystack.substr(0, start) + haystack.substr(start).replace(regExp, replacement);
   const index = occurrence - 1;
   let relevantString = haystack.substring(start);
   let a, count = 0;

diff --git a/udfs/community/test_cases.js b/udfs/community/test_cases.js
@@ -1453,6 +1453,24 @@ generate_udf_test("cw_regexp_replace_4", [
     expected_output: `"TestCad$123456"`,
   },
 ]);
+generate_udf_test("cw_regexp_replace_4", [
+  {
+    inputs: [`"TestStr123456Str"`, `"Str"`, `"Cad$"`, `CAST(1 AS INT64)`],
+    expected_output: `"TestCad$123456Cad$"`,
+  },
+]);
+generate_udf_test("cw_regexp_replace_4", [
+  {
+    inputs: [`"abaa"`, `"a"`, `"A"`, `CAST(1 AS INT64)`],
+    expected_output: `"AbAA"`,
+  },
+]);
+generate_udf_test("cw_regexp_replace_4", [
+  {
+    inputs: [`"abaa"`, `"a"`, `"A"`, `CAST(2 AS INT64)`],
+    expected_output: `"abAA"`,
+  },
+]);
 generate_udf_test("cw_regexp_replace_5", [
   {
     inputs: [