add encodeXMLComponent function

src/Functions/encodeXML.cpp

@@ -0,0 +1,142 @@
+#include <Columns/ColumnString.h>
+#include <Functions/FunctionFactory.h>
+#include <Functions/FunctionStringToString.h>
+#include <common/find_symbols.h>
+
+namespace DB
+{
+namespace ErrorCodes
+{
+    extern const int ILLEGAL_TYPE_OF_ARGUMENT;
+}
+
+namespace
+{
+    struct EncodeXMLName
+    {
+        static constexpr auto name = "encodeXMLComponent";
+    };
+
+    class FunctionEncodeXMLImpl
+    {
+    public:
+        static void vector(
+            const ColumnString::Chars & data,
+            const ColumnString::Offsets & offsets,
+            ColumnString::Chars & res_data,
+            ColumnString::Offsets & res_offsets)
+        {
+            res_data.resize(data.size());
+            size_t size = offsets.size();
+            res_offsets.resize(size);
+
+            size_t prev_offset = 0;
+            size_t res_offset = 0;
+
+            for (size_t i = 0; i < size; ++i)
+            {
+                const char * src_data = reinterpret_cast<const char *>(&data[prev_offset]);
+                size_t src_size = offsets[i] - prev_offset;
+                size_t dst_size = execute(src_data, src_size, reinterpret_cast<char *>(res_data.data() + res_offset));
+
+                res_offset += dst_size;
+                res_offsets[i] = res_offset;
+                prev_offset = offsets[i];
+            }
+
+            res_data.resize(res_offset);
+        }
+
+        static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &)
+        {
+            throw Exception("Function encodeXML cannot work with FixedString argument", ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
+        }
+
+    private:
+        static size_t execute(const char * src, size_t src_size, char * dst)
+        {
+            const char * src_prev_pos = src;
+            const char * src_curr_pos = src;
+            const char * src_end = src + src_size;
+            char * dst_pos = dst;
+
+            while (true)
+            {
+                src_curr_pos = find_first_symbols<'<', '&', '>', '"', '\''>(src_curr_pos, src_end);
+
+                if (src_curr_pos == src_end)
+                {
+                    break;
+                }
+                else if (*src_curr_pos == '<')
+                {
+                    size_t bytes_to_copy = src_curr_pos - src_prev_pos;
+                    memcpySmallAllowReadWriteOverflow15(dst_pos, src_prev_pos, bytes_to_copy);
+                    dst_pos += bytes_to_copy;
+                    memcpySmallAllowReadWriteOverflow15(dst_pos, "&lt;", 4);
+                    dst_pos += 4;
+                    src_prev_pos = src_curr_pos + 1;
+                    src_curr_pos++;
+                }
+                else if (*src_curr_pos == '&')
+                {
+                    size_t bytes_to_copy = src_curr_pos - src_prev_pos;
+                    memcpySmallAllowReadWriteOverflow15(dst_pos, src_prev_pos, bytes_to_copy);
+                    dst_pos += bytes_to_copy;
+                    memcpySmallAllowReadWriteOverflow15(dst_pos, "&amp;", 5);
+                    dst_pos += 5;
+                    src_prev_pos = src_curr_pos + 1;
+                    src_curr_pos++;
+                }
+                else if (*src_curr_pos == '>')
+                {
+                    size_t bytes_to_copy = src_curr_pos - src_prev_pos;
+                    memcpySmallAllowReadWriteOverflow15(dst_pos, src_prev_pos, bytes_to_copy);
+                    dst_pos += bytes_to_copy;
+                    memcpySmallAllowReadWriteOverflow15(dst_pos, "&gt;", 4);
+                    dst_pos += 4;
+                    src_prev_pos = src_curr_pos + 1;
+                    src_curr_pos++;
+                }
+                else if (*src_curr_pos == '"')
+                {
+                    size_t bytes_to_copy = src_curr_pos - src_prev_pos;
+                    memcpySmallAllowReadWriteOverflow15(dst_pos, src_prev_pos, bytes_to_copy);
+                    dst_pos += bytes_to_copy;
+                    memcpySmallAllowReadWriteOverflow15(dst_pos, "&quot;", 6);
+                    dst_pos += 6;
+                    src_prev_pos = src_curr_pos + 1;
+                    src_curr_pos++;
+                }
+                else if (*src_curr_pos == '\'')
+                {
+                    size_t bytes_to_copy = src_curr_pos - src_prev_pos;
+                    memcpySmallAllowReadWriteOverflow15(dst_pos, src_prev_pos, bytes_to_copy);
+                    dst_pos += bytes_to_copy;
+                    memcpySmallAllowReadWriteOverflow15(dst_pos, "&apos;", 6);
+                    dst_pos += 6;
+                    src_prev_pos = src_curr_pos + 1;
+                    src_curr_pos++;
+                }
+            }
+
+            if (src_prev_pos < src_curr_pos)
+            {
+                size_t bytes_to_copy = src_curr_pos - src_prev_pos;
+                memcpySmallAllowReadWriteOverflow15(dst_pos, src_prev_pos, bytes_to_copy);
+                dst_pos += bytes_to_copy;
+            }
+
+            return dst_pos - dst;
+        }
+    };
+
+    using FunctionEncodeXML = FunctionStringToString<FunctionEncodeXMLImpl, EncodeXMLName>;
+
+}
+
+void registerFunctionEncodeXML(FunctionFactory & factory)
+{
+    factory.registerFunction<FunctionEncodeXML>();
+}
+}

src/Functions/registerFunctionsString.cpp

@@ -33,6 +33,7 @@ void registerFunctionRegexpQuoteMeta(FunctionFactory &);
 void registerFunctionNormalizeQuery(FunctionFactory &);
 void registerFunctionNormalizedQueryHash(FunctionFactory &);
 void registerFunctionCountMatches(FunctionFactory &);
+void registerFunctionEncodeXML(FunctionFactory & factory);
 
 #if USE_BASE64
 void registerFunctionBase64Encode(FunctionFactory &);
@@ -68,6 +69,7 @@ void registerFunctionsString(FunctionFactory & factory)
     registerFunctionNormalizeQuery(factory);
     registerFunctionNormalizedQueryHash(factory);
     registerFunctionCountMatches(factory);
+    registerFunctionEncodeXML(factory);
 #if USE_BASE64
     registerFunctionBase64Encode(factory);
     registerFunctionBase64Decode(factory);

src/Functions/ya.make

@@ -224,6 +224,7 @@ SRCS(
     dumpColumnStructure.cpp
     e.cpp
     empty.cpp
+    encodeXML.cpp
     encrypt.cpp
     endsWith.cpp
     equals.cpp

tests/queries/0_stateless/01600_encode_XML.reference

@@ -0,0 +1,4 @@
+Hello, "world"!
+<123>
+&clickhouse
+'foo'

tests/queries/0_stateless/01600_encode_XML.sql

@@ -0,0 +1,4 @@
+SELECT encodeXMLComponent('Hello, "world"!');
+SELECT encodeXMLComponent('<123>');
+SELECT encodeXMLComponent('&clickhouse');
+SELECT encodeXMLComponent('\'foo\'');