Extend Split to work with multiple separators

martinus · martinus · commit b7ab9db54549 · 2022-05-04T07:34:47.000+02:00
diff --git a/src/test/util_tests.cpp b/src/test/util_tests.cpp
@@ -2396,6 +2396,19 @@ BOOST_AUTO_TEST_CASE(test_SplitString)
         BOOST_CHECK_EQUAL(result.size(), 1);
         BOOST_CHECK_EQUAL(result[0], "AAA");
     }
+
+    // multiple split characters
+    {
+        using V = std::vector<std::string>;
+        BOOST_TEST(SplitString("a,b.c:d;e", ",;") == V({"a", "b.c:d", "e"}));
+        BOOST_TEST(SplitString("a,b.c:d;e", ",;:.") == V({"a", "b", "c", "d", "e"}));
+        BOOST_TEST(SplitString("a,b.c:d;e", "") == V({"a,b.c:d;e"}));
+        BOOST_TEST(SplitString("aaa", "bcdefg") == V({"aaa"}));
+        BOOST_TEST(SplitString("x\0a,b"s, "\0"s) == V({"x", "a,b"}));
+        BOOST_TEST(SplitString("x\0a,b"s, '\0') == V({"x", "a,b"}));
+        BOOST_TEST(SplitString("x\0a,b"s, "\0,"s) == V({"x", "a", "b"}));
+        BOOST_TEST(SplitString("abcdefg", "bcd") == V({"a", "", "", "efg"}));
+    }
 }
 
 BOOST_AUTO_TEST_CASE(test_LogEscapeMessage)
diff --git a/src/util/spanparsing.h b/src/util/spanparsing.h
@@ -8,6 +8,7 @@
 #include <span.h>
 
 #include <string>
+#include <string_view>
 #include <vector>
 
 namespace spanparsing {
@@ -36,21 +37,21 @@ bool Func(const std::string& str, Span<const char>& sp);
  */
 Span<const char> Expr(Span<const char>& sp);
 
-/** Split a string on every instance of sep, returning a vector.
+/** Split a string on any char found in separators, returning a vector.
  *
  * If sep does not occur in sp, a singleton with the entirety of sp is returned.
  *
  * Note that this function does not care about braces, so splitting
  * "foo(bar(1),2),3) on ',' will return {"foo(bar(1)", "2)", "3)"}.
  */
 template <typename T = Span<const char>>
-std::vector<T> Split(const Span<const char>& sp, char sep)
+std::vector<T> Split(const Span<const char>& sp, std::string_view separators)
 {
     std::vector<T> ret;
     auto it = sp.begin();
     auto start = it;
     while (it != sp.end()) {
-        if (*it == sep) {
+        if (separators.find(*it) != std::string::npos) {
             ret.emplace_back(start, it);
             start = it + 1;
         }
@@ -60,6 +61,19 @@ std::vector<T> Split(const Span<const char>& sp, char sep)
     return ret;
 }
 
+/** Split a string on every instance of sep, returning a vector.
+ *
+ * If sep does not occur in sp, a singleton with the entirety of sp is returned.
+ *
+ * Note that this function does not care about braces, so splitting
+ * "foo(bar(1),2),3) on ',' will return {"foo(bar(1)", "2)", "3)"}.
+ */
+template <typename T = Span<const char>>
+std::vector<T> Split(const Span<const char>& sp, char sep)
+{
+    return Split<T>(sp, std::string_view{&sep, 1});
+}
+
 } // namespace spanparsing
 
 #endif // BITCOIN_UTIL_SPANPARSING_H
diff --git a/src/util/string.h b/src/util/string.h
@@ -14,13 +14,19 @@
 #include <locale>
 #include <sstream>
 #include <string>
+#include <string_view>
 #include <vector>
 
 [[nodiscard]] inline std::vector<std::string> SplitString(std::string_view str, char sep)
 {
     return spanparsing::Split<std::string>(str, sep);
 }
 
+[[nodiscard]] inline std::vector<std::string> SplitString(std::string_view str, std::string_view separators)
+{
+    return spanparsing::Split<std::string>(str, separators);
+}
+
 [[nodiscard]] inline std::string_view TrimStringView(std::string_view str, std::string_view pattern = " \f\n\r\t\v")
 {
     std::string::size_type front = str.find_first_not_of(pattern);

Original file line number	Diff line number	Diff line change
`@@ -14,13 +14,19 @@`
`14`	`14`	`#include <locale>`
`15`	`15`	`#include <sstream>`
`16`	`16`	`#include <string>`
	`17`	`+#include <string_view>`
`17`	`18`	`#include <vector>`
`18`	`19`
`19`	`20`	`[[nodiscard]] inline std::vector<std::string> SplitString(std::string_view str, char sep)`
`20`	`21`	`{`
`21`	`22`	`return spanparsing::Split<std::string>(str, sep);`
`22`	`23`	`}`
`23`	`24`
	`25`	`+[[nodiscard]] inline std::vector<std::string> SplitString(std::string_view str, std::string_view separators)`
	`26`	`+{`
	`27`	`+ return spanparsing::Split<std::string>(str, separators);`
	`28`	`+}`
	`29`	`+`
`24`	`30`	`[[nodiscard]] inline std::string_view TrimStringView(std::string_view str, std::string_view pattern = " \f\n\r\t\v")`
`25`	`31`	`{`
`26`	`32`	`std::string::size_type front = str.find_first_not_of(pattern);`