Backport #32138 to 21.12: Fix 'APPLY lambda' parsing

docker/test/fuzzer/run-fuzzer.sh

@@ -189,6 +189,7 @@ continue
         --receive_data_timeout_ms=10000 \
         --stacktrace \
         --query-fuzzer-runs=1000 \
+        --testmode \
         --queries-file $(ls -1 ch/tests/queries/0_stateless/*.sql | sort -R) \
         $NEW_TESTS_OPT \
         > >(tail -n 100000 > fuzzer.log) \

programs/client/Client.cpp

@@ -705,16 +705,16 @@ bool Client::processWithFuzzing(const String & full_query)
             throw;
     }
 
-    // `USE db` should not be executed
-    // since this will break every query after `DROP db`
-    if (orig_ast->as<ASTUseQuery>())
+    if (!orig_ast)
     {
+        // Can't continue after a parsing error
         return true;
     }
 
-    if (!orig_ast)
+    // `USE db` should not be executed
+    // since this will break every query after `DROP db`
+    if (orig_ast->as<ASTUseQuery>())
     {
-        // Can't continue after a parsing error
         return true;
     }
 

src/Parsers/ExpressionElementParsers.cpp

@@ -22,7 +22,6 @@
 #include <Parsers/ASTSubquery.h>
 #include <Parsers/ASTTTLElement.h>
 #include <Parsers/ASTWindowDefinition.h>
-#include <Parsers/IAST.h>
 #include <Parsers/ASTAssignment.h>
 
 #include <Parsers/parseIdentifierOrStringLiteral.h>
@@ -35,7 +34,6 @@
 #include <Parsers/ParserCreateQuery.h>
 
 #include <Parsers/queryToString.h>
-#include <boost/algorithm/string.hpp>
 #include "ASTColumnsMatcher.h"
 
 #include <Interpreters/StorageID.h>
@@ -1935,15 +1933,21 @@ bool ParserColumnsTransformers::parseImpl(Pos & pos, ASTPtr & node, Expected & e
         {
             if (const auto * func = lambda->as<ASTFunction>(); func && func->name == "lambda")
             {
+                if (func->arguments->children.size() != 2)
+                    throw Exception(ErrorCodes::SYNTAX_ERROR, "lambda requires two arguments");
+
                 const auto * lambda_args_tuple = func->arguments->children.at(0)->as<ASTFunction>();
+                if (!lambda_args_tuple || lambda_args_tuple->name != "tuple")
+                    throw Exception(ErrorCodes::SYNTAX_ERROR, "First argument of lambda must be a tuple");
+
                 const ASTs & lambda_arg_asts = lambda_args_tuple->arguments->children;
                 if (lambda_arg_asts.size() != 1)
-                    throw Exception(ErrorCodes::BAD_ARGUMENTS, "APPLY column transformer can only accept lambda with one argument");
+                    throw Exception(ErrorCodes::SYNTAX_ERROR, "APPLY column transformer can only accept lambda with one argument");
 
                 if (auto opt_arg_name = tryGetIdentifierName(lambda_arg_asts[0]); opt_arg_name)
                     lambda_arg = *opt_arg_name;
                 else
-                    throw Exception(ErrorCodes::BAD_ARGUMENTS, "lambda argument declarations must be identifiers");
+                    throw Exception(ErrorCodes::SYNTAX_ERROR, "lambda argument declarations must be identifiers");
             }
             else
             {

tests/queries/0_stateless/02128_apply_lambda_parsing.reference

@@ -0,0 +1,10 @@
+1
+1
+1
+1
+1
+1
+2
+3
+4
+5

tests/queries/0_stateless/02128_apply_lambda_parsing.sql

@@ -0,0 +1,13 @@
+WITH * APPLY lambda(e); -- { clientError SYNTAX_ERROR }
+SELECT * APPLY lambda(); -- { clientError SYNTAX_ERROR }
+SELECT * APPLY lambda(1); -- { clientError SYNTAX_ERROR }
+SELECT * APPLY lambda(x); -- { clientError SYNTAX_ERROR }
+SELECT * APPLY lambda(range(1)); -- { clientError SYNTAX_ERROR }
+SELECT * APPLY lambda(range(x)); -- { clientError SYNTAX_ERROR }
+SELECT * APPLY lambda(1, 2); -- { clientError SYNTAX_ERROR }
+SELECT * APPLY lambda(x, y); -- { clientError SYNTAX_ERROR }
+SELECT * APPLY lambda((x, y), 2); -- { clientError SYNTAX_ERROR }
+SELECT * APPLY lambda((x, y), x + y); -- { clientError SYNTAX_ERROR }
+SELECT * APPLY lambda(tuple(1), 1); -- { clientError SYNTAX_ERROR }
+SELECT * APPLY lambda(tuple(x), 1) FROM numbers(5);
+SELECT * APPLY lambda(tuple(x), x + 1) FROM numbers(5);