fix: clustering_information panic if string domain is none

src/query/storages/fuse/src/table_functions/clustering_information.rs

@@ -296,7 +296,14 @@ impl<'a> ClusteringInformation<'a> {
         let (keys, values): (Vec<_>, Vec<_>) = points_map.into_iter().unzip();
         let cluster_key_types = exprs
             .into_iter()
-            .map(|v| v.data_type().clone())
+            .map(|v| {
+                let data_type = v.data_type();
+                if matches!(*data_type, DataType::String) {
+                    data_type.wrap_nullable()
+                } else {
+                    data_type.clone()
+                }
+            })
             .collect::<Vec<_>>();
         let indices = compare_scalars(keys, &cluster_key_types)?;
         for idx in indices.into_iter() {

tests/sqllogictests/suites/base/09_fuse_engine/09_0014_func_clustering_information_function.test

@@ -86,4 +86,27 @@ select * exclude(timestamp) from clustering_information('default','t09_0014_1',
 (SUBSTRING(c FROM 2 FOR 2)) linear 2 0 1.0 2.0 {"00002":2}
 
 statement ok
-drop table t09_0014_1
+drop table t09_0014_1 all
+
+##issue https://github.com/databendlabs/databend/issues/16763
+statement ok
+create table t09_0014_2(c varchar not null) cluster by(substring(c FROM 1 FOR 4))
+
+statement ok
+insert into t09_0014_2 values('abcde'),('bcdef')
+
+statement ok
+insert into t09_0014_2 values('bcdff'),('cdefg')
+
+query TTIIFFT
+select * exclude(timestamp) from clustering_information('default','t09_0014_2')
+----
+(SUBSTRING(c FROM 1 FOR 4)) linear 2 0 0.0 1.0 {"00001":2}
+
+query TTIIFFT
+select * exclude(timestamp) from clustering_information('default','t09_0014_2', 'substr(c,2,4)')
+----
+(SUBSTRING(c FROM 2 FOR 4)) linear 2 0 1.0 2.0 {"00002":2}
+
+statement ok
+drop table t09_0014_2 all

tests/suites/1_stateful/09_http_handler/09_0007_token.py

@@ -133,10 +133,10 @@ def fake_expired_token(ty):
         "sid": "",
     }
     return (
-            "bend-v1-"
-            + ty
-            + "-"
-            + base64.b64encode(json.dumps(expired_claim).encode("utf-8")).decode("utf-8")
+        "bend-v1-"
+        + ty
+        + "-"
+        + base64.b64encode(json.dumps(expired_claim).encode("utf-8")).decode("utf-8")
     )
 
 
@@ -207,6 +207,7 @@ def main():
 
 if __name__ == "__main__":
     import logging
+
     try:
         main()
     except Exception as e:

tests/suites/1_stateful/09_http_handler/09_0009_cookie.py

@@ -8,25 +8,26 @@
 auth = ("root", "")
 logging.basicConfig(level=logging.ERROR, format="%(asctime)s %(levelname)s %(message)s")
 
+
 class GlobalCookieJar(RequestsCookieJar):
     def __init__(self):
         super().__init__()
 
     def set_cookie(self, cookie: Cookie, *args, **kwargs):
-        cookie.domain = ''
-        cookie.path = '/'
+        cookie.domain = ""
+        cookie.path = "/"
         super().set_cookie(cookie, *args, **kwargs)
 
     def get_dict(self, domain=None, path=None):
         # 忽略 domain 和 path 参数，返回所有 Cookie
         return {cookie.name: cookie.value for cookie in self}
 
 
-def do_query(session_client, query, session_state=None ):
+def do_query(session_client, query, session_state=None):
     url = f"http://localhost:8000/v1/query"
     query_payload = {
         "sql": query,
-        "pagination": {"wait_time_secs": 100, 'max_rows_per_page': 2},
+        "pagination": {"wait_time_secs": 100, "max_rows_per_page": 2},
     }
     if session_state:
         query_payload["session"] = session_state
@@ -47,10 +48,10 @@ def test_simple():
     assert resp.status_code == 200, resp.text
     assert resp.json()["data"] == [["1"]], resp.text
     sid = client.cookies.get("session_id")
-    #print(sid)
+    # print(sid)
 
     last_access_time1 = int(client.cookies.get("last_access_time"))
-    #print(last_access_time1)
+    # print(last_access_time1)
     assert time.time() - 10 < last_access_time1 < time.time()
 
     time.sleep(1.5)
@@ -83,7 +84,7 @@ def test_temp_table():
 
     resp = do_query(client, "select * from t1", session_state)
     assert resp.status_code == 200, resp.text
-    assert resp.json()['data'] == [["3"], ["4"]]
+    assert resp.json()["data"] == [["3"], ["4"]]
     session_state = resp.json()["session"]
     assert session_state["need_sticky"], resp.text
     assert session_state["need_keep_alive"]
@@ -99,6 +100,7 @@ def main():
     test_simple()
     test_temp_table()
 
+
 if __name__ == "__main__":
     try:
         main()