Merge pull request #54516 from ClickHouse/backport/23.8/54373

Backport #54373 to 23.8: Fix possible error 'URI contains invalid characters' in s3 table function
ClickHouse · Sep 12, 2023 · f273f04 · f273f04
2 parents 311d53e + a2adaf1
commit f273f04
Show file tree

Hide file tree

Showing 3 changed files with 16 additions and 4 deletions.
diff --git a/src/TableFunctions/TableFunctionS3.cpp b/src/TableFunctions/TableFunctionS3.cpp
@@ -58,6 +58,11 @@ void TableFunctionS3::parseArgumentsImpl(ASTs & args, const ContextPtr & context
     if (auto named_collection = tryGetNamedCollectionWithOverrides(args, context))
     {
         StorageS3::processNamedCollectionResult(configuration, *named_collection);
+        if (configuration.format == "auto")
+        {
+            String file_path = named_collection->getOrDefault<String>("filename", Poco::URI(named_collection->get<String>("url")).getPath());
+            configuration.format = FormatFactory::instance().getFormatFromFileName(file_path, true);
+        }
     }
     else
     {
@@ -152,7 +157,8 @@ void TableFunctionS3::parseArgumentsImpl(ASTs & args, const ContextPtr & context
         }
 
         /// This argument is always the first
-        configuration.url = S3::URI(checkAndGetLiteralArgument<String>(args[0], "url"));
+        String url = checkAndGetLiteralArgument<String>(args[0], "url");
+        configuration.url = S3::URI(url);
 
         if (args_to_idx.contains("format"))
         {
@@ -176,12 +182,12 @@ void TableFunctionS3::parseArgumentsImpl(ASTs & args, const ContextPtr & context
             configuration.auth_settings.secret_access_key = checkAndGetLiteralArgument<String>(args[args_to_idx["secret_access_key"]], "secret_access_key");
 
         configuration.auth_settings.no_sign_request = no_sign_request;
+
+        if (configuration.format == "auto")
+            configuration.format = FormatFactory::instance().getFormatFromFileName(Poco::URI(url).getPath(), true);
     }
 
     configuration.keys = {configuration.url.key};
-
-    if (configuration.format == "auto")
-        configuration.format = FormatFactory::instance().getFormatFromFileName(Poco::URI(configuration.url.uri.getPath()).getPath(), true);
 }
 
 void TableFunctionS3::parseArguments(const ASTPtr & ast_function, ContextPtr context)

diff --git a/tests/queries/0_stateless/02873_s3_presigned_url_and_url_with_special_characters.reference b/tests/queries/0_stateless/02873_s3_presigned_url_and_url_with_special_characters.reference
diff --git a/tests/queries/0_stateless/02873_s3_presigned_url_and_url_with_special_characters.sql b/tests/queries/0_stateless/02873_s3_presigned_url_and_url_with_special_characters.sql
@@ -0,0 +1,6 @@
+-- Tags: no-fasttest
+
+select * from s3('https://datasets-documentation.s3.eu-west-3.amazonaws.com/MyPrefix/BU%20-%20UNIT%20-%201/*.parquet'); -- { serverError CANNOT_EXTRACT_TABLE_STRUCTURE }
+
+select * from s3('https://datasets-documentation.s3.eu-west-3.amazonaws.com/MyPrefix/*.parquet?some_tocken=ABCD'); -- { serverError CANNOT_EXTRACT_TABLE_STRUCTURE }
+