fix and add test for storage file version

dbms/src/Common/parseGlobs.cpp

@@ -5,7 +5,10 @@
 
 namespace DB
 {
-
+/*  Because of difference between grep-wildcard-syntax and perl-regexp one we need some transformation of string to use RE2 library for matching.
+ *  It couldn't be one pass because of various configurations of braces in filenames (Linux allow almost any symbols in paths).
+ *  So there are some iterations of escaping and replacements to make correct perl-regexp.
+ */
 std::string makeRegexpPatternFromGlobs(const std::string & initial_str)
 {
     std::string first_prepare;

dbms/src/Storages/StorageFile.cpp

@@ -63,7 +63,7 @@ std::vector<std::string> LSWithRegexpMatching(const std::string & path_for_ls, c
 
     std::vector<std::string> result;
     fs::directory_iterator end;
-    for (fs::directory_iterator it(path_for_ls); it != end; ++it)
+    for (fs::directory_iterator it(path_for_ls + for_match.substr(1, end_of_path_without_globs)); it != end; ++it)
     {
         std::string full_path = it->path().string();
         size_t last_slash = full_path.rfind('/');
@@ -139,7 +139,7 @@ StorageFile::StorageFile(
             if (first_glob != std::string::npos)
             {
                 path_with_globs = true;
-                matched_paths = LSWithRegexpMatching(db_dir_path, path);
+                matched_paths = LSWithRegexpMatching("/", path);
                 for (const auto & cur_path : matched_paths)
                     checkCreationIsAllowed(context_global, db_dir_path, cur_path, table_fd);
             }

dbms/tests/integration/test_globs_in_filepath/test.py

@@ -0,0 +1,42 @@
+import pytest
+
+from helpers.cluster import ClickHouseCluster
+
+cluster = ClickHouseCluster(__file__)
+node = cluster.add_instance('node')
+
+@pytest.fixture(scope="module")
+def start_cluster():
+    try:
+        cluster.start()
+        yield cluster
+    finally:
+        cluster.shutdown()
+
+def test_globs(start_cluster):
+    some_data = "\t555.222\nData\t777.333"
+    path_to_userfiles = "/var/lib/clickhouse/user_files/"
+    dirs = ["dir1/", "dir2/"]
+
+    for dir in dirs:
+        node.exec_in_container(['bash', '-c', 'mkdir {}{}'.format(path_to_userfiles, dir)], privileged=True, user='root')
+
+    # all directories appeared in files must be listed in dirs
+    files = ["dir1/file1", "dir1/file2",
+             "dir2/file1", "dir2/file2", "dir2/file11",
+             "file1"]
+
+    for filename in files:
+        node.exec_in_container(['bash', '-c', 'echo "{}{}" > {}{}'.format(filename, some_data, path_to_userfiles, filename)], privileged=True, user='root')
+
+    test_requests = [("dir1/file*", "4"),
+                     ("dir1/file?", "4"),
+                     ("dir1/file{0..9}", "4"),
+                     ("dir2/file*", "6"),
+                     ("dir2/file?", "4"),
+                     ("*", "2")]
+
+    for pattern, value in test_requests:
+        assert node.query('''
+            select count(*) from file('{}', 'TSV', 'text String, number Float64')
+        '''.format(pattern)) == '{}\n'.format(value)