Add port() function

It uses domain() (getURLHost()) internally so it has the same
limitations:
- no ipv6 support
- no host-without-dots support (i.e. "foobar")

docs/en/sql-reference/functions/url-functions.md

@@ -117,6 +117,10 @@ Returns the part of the domain that includes top-level subdomains up to the “f
 
 For example, `cutToFirstSignificantSubdomain('https://news.yandex.com.tr/') = 'yandex.com.tr'`.
 
+### port(URL) {#port}
+
+Returns the port or zero if there is no port in the URL (or in case of validation error).
+
 ### path {#path}
 
 Returns the path. Example: `/top/news.html` The path does not include the query string.

src/Functions/URL/port.cpp

@@ -0,0 +1,95 @@
+#include <Functions/FunctionFactory.h>
+#include <Functions/FunctionStringOrArrayToT.h>
+#include <Common/StringUtils/StringUtils.h>
+#include "domain.h"
+
+
+namespace DB
+{
+
+namespace ErrorCodes
+{
+    extern const int ILLEGAL_TYPE_OF_ARGUMENT;
+}
+
+struct ExtractPort
+{
+    static constexpr auto name = "port";
+    static constexpr auto is_fixed_to_constant = true;
+
+    static void vector(const ColumnString::Chars & data, const ColumnString::Offsets & offsets, PaddedPODArray<UInt16> & res)
+    {
+        size_t size = offsets.size();
+
+        ColumnString::Offset prev_offset = 0;
+        for (size_t i = 0; i < size; ++i)
+        {
+            res[i] = parse(data, prev_offset, offsets[i] - prev_offset - 1);
+            prev_offset = offsets[i];
+        }
+    }
+
+    static void vectorFixedToConstant(const ColumnString::Chars & data, size_t n, UInt16 & res) { res = parse(data, 0, n); }
+
+    static void vectorFixedToVector(const ColumnString::Chars & data, size_t n, PaddedPODArray<UInt16> & res)
+    {
+        size_t size = data.size() / n;
+
+        for (size_t i = 0; i < size; ++i)
+        {
+            res[i] = parse(data, i * n, n);
+        }
+    }
+
+    [[noreturn]] static void array(const ColumnString::Offsets &, PaddedPODArray<UInt16> &)
+    {
+        throw Exception("Cannot apply function port to Array argument", ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
+    }
+
+private:
+    static UInt16 parse(const ColumnString::Chars & buf, size_t offset, size_t size)
+    {
+        const char * p = reinterpret_cast<const char *>(&buf[0]) + offset;
+        const char * end = p + size;
+
+        StringRef host = getURLHost(p, size);
+        if (!host.size)
+            return 0;
+        if (host.size == size)
+            return 0;
+
+        p = host.data + host.size;
+        if (*p++ != ':')
+            return 0;
+
+        Int64 port = 0;
+        while (p < end)
+        {
+            if (*p == '/')
+                break;
+            if (!isNumericASCII(*p))
+                return 0;
+
+            port = (port * 10) + (*p - '0');
+            if (port < 0 || port > UInt16(-1))
+                return 0;
+            ++p;
+        }
+        return port;
+    }
+};
+
+struct NamePort
+{
+    static constexpr auto name = "port";
+};
+
+using FunctionPort = FunctionStringOrArrayToT<ExtractPort, NamePort, UInt16>;
+
+void registerFunctionPort(FunctionFactory & factory)
+{
+    factory.registerFunction<FunctionPort>();
+}
+
+}
+

src/Functions/URL/registerFunctionsURL.cpp

@@ -8,6 +8,7 @@ void registerFunctionDomain(FunctionFactory & factory);
 void registerFunctionDomainWithoutWWW(FunctionFactory & factory);
 void registerFunctionFirstSignificantSubdomain(FunctionFactory & factory);
 void registerFunctionTopLevelDomain(FunctionFactory & factory);
+void registerFunctionPort(FunctionFactory & factory);
 void registerFunctionPath(FunctionFactory & factory);
 void registerFunctionPathFull(FunctionFactory & factory);
 void registerFunctionQueryString(FunctionFactory & factory);
@@ -33,6 +34,7 @@ void registerFunctionsURL(FunctionFactory & factory)
     registerFunctionDomainWithoutWWW(factory);
     registerFunctionFirstSignificantSubdomain(factory);
     registerFunctionTopLevelDomain(factory);
+    registerFunctionPort(factory);
     registerFunctionPath(factory);
     registerFunctionPathFull(factory);
     registerFunctionQueryString(factory);

tests/queries/0_stateless/01284_port.reference

@@ -0,0 +1,21 @@
+ipv4
+0
+80
+80
+80
+80
+hostname
+0
+80
+80
+80
+80
+ipv6
+0
+0
+0
+0
+0
+0
+host-no-dot
+0

tests/queries/0_stateless/01284_port.sql

@@ -0,0 +1,25 @@
+select 'ipv4';
+select port('http://127.0.0.1/');
+select port('http://127.0.0.1:80');
+select port('http://127.0.0.1:80/');
+select port('//127.0.0.1:80/');
+select port('127.0.0.1:80');
+select 'hostname';
+select port('http://foobar.com/');
+select port('http://foobar.com:80');
+select port('http://foobar.com:80/');
+select port('//foobar.com:80/');
+select port('foobar.com:80');
+
+--
+-- Known limitations of domain() (getURLHost())
+--
+select 'ipv6';
+select port('http://[2001:db8::8a2e:370:7334]/');
+select port('http://[2001:db8::8a2e:370:7334]:80');
+select port('http://[2001:db8::8a2e:370:7334]:80/');
+select port('//[2001:db8::8a2e:370:7334]:80/');
+select port('[2001:db8::8a2e:370:7334]:80');
+select port('2001:db8::8a2e:370:7334:80');
+select 'host-no-dot';
+select port('//foobar:80/');