refine unittest

Author: zhichao-li

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/stringOperations.scala

@@ -434,7 +434,7 @@ case class Substring_index(strExpr: Expression, delimExpr: Expression, countExpr
     } else {
       val idx = lastOrdinalIndexOf(strUtf8, delimUtf8, -count)
       if (idx != -1) {
-        strUtf8.substring(idx + 1, strUtf8.numChars())
+        strUtf8.substring(idx + delimUtf8.numChars(), strUtf8.numChars())
       } else {
         strUtf8
       }

sql/core/src/test/scala/org/apache/spark/sql/StringFunctionsSuite.scala

@@ -157,33 +157,60 @@ class StringFunctionsSuite extends QueryTest {
   }
 
   test("string substring_index function") {
-    val df = Seq(("ac,ab,ad,ab,cc", "aa", "zz")).toDF("a", "b", "c")
+    val df = Seq(("www.apache.org", ".", "zz")).toDF("a", "b", "c")
     checkAnswer(
-      df.select(substring_index($"a", ",", 2)),
-      Row("ac,ab"))
+      df.select(substring_index($"a", ".", 3)),
+      Row("www.apache.org"))
     checkAnswer(
-      df.select(substring_index($"a", "ab", 2)),
-      Row("ac,ab,ad,"))
+      df.select(substring_index($"a", ".", 2)),
+      Row("www.apache"))
     checkAnswer(
-      df.select(substring_index(lit(""), "ab", 2)),
+      df.select(substring_index($"a", ".", 1)),
+      Row("www"))
+    checkAnswer(
+      df.select(substring_index($"a", ".", 0)),
+      Row(""))
+    checkAnswer(
+      df.select(substring_index(lit("www.apache.org"), ".", -1)),
+      Row("org"))
+    checkAnswer(
+      df.select(substring_index(lit("www.apache.org"), ".", -2)),
+      Row("apache.org"))
+    checkAnswer(
+      df.select(substring_index(lit("www.apache.org"), ".", -3)),
+      Row("www.apache.org"))
+    // str is empty string
+    checkAnswer(
+      df.select(substring_index(lit(""), ".", 1)),
+      Row(""))
+    // empty string delim
+    checkAnswer(
+      df.select(substring_index(lit("www.apache.org"), "", 1)),
       Row(""))
+    // delim does not exist in str
     checkAnswer(
-      df.select(substring_index(lit(null), "ab", 2)),
+      df.select(substring_index(lit("www.apache.org"), "#", 1)),
+      Row("www.apache.org"))
+    // delim is 2 chars
+    checkAnswer(
+      df.select(substring_index(lit("www||apache||org"), "||", 2)),
+      Row("www||apache"))
+    checkAnswer(
+      df.select(substring_index(lit("www||apache||org"), "||", -2)),
+      Row("apache||org"))
+    // null
+    checkAnswer(
+      df.select(substring_index(lit(null), "||", 2)),
+      Row(null))
+    checkAnswer(
+      df.select(substring_index(lit("www.apache.org"), null, 2)),
       Row(null))
+    // non ascii chars
     // scalastyle:off
     checkAnswer(
-      df.select(substring_index(lit("大千世界大千世界"), "千", 2)),
+      df.selectExpr("""substring_index("大千世界大千世界", "千", 2)"""),
       Row("大千世界大"))
     // scalastyle:on
-    checkAnswer(
-      df.selectExpr("""substring_index(a, ",", 2)"""),
-      Row("ac,ab"))
-    checkAnswer(
-      df.selectExpr("""substring_index(a, ",", -2)"""),
-      Row("ab,cc"))
-    checkAnswer(
-      df.selectExpr("""substring_index(a, ",", 10)"""),
-      Row("ac,ab,ad,ab,cc"))
   }
 
   test("string locate function") {

unsafe/src/main/java/org/apache/spark/unsafe/types/UTF8String.java

@@ -381,7 +381,7 @@ private int firstOfCurrentCodePoint(int bytePos) {
     throw new RuntimeException("Invalid utf8 string");
   }
 
-  private int endByte(int startCodePoint) {
+  private int indexEnd(int startCodePoint) {
     int i = numBytes -1; // position in byte
     int c = numChars() - 1; // position in character
     while (i >=0 && c > startCodePoint) {
@@ -398,7 +398,7 @@ public int lastIndexOf(UTF8String v, int startCodePoint) {
     if (numBytes == 0) {
       return -1;
     }
-    int fromIndexEnd = endByte(startCodePoint);
+    int fromIndexEnd = indexEnd(startCodePoint);
     int count = startCodePoint;
     int vNumChars = v.numChars();
     do {