Fix describe throwing errors when listing

src/databricks/labs/ucx/hive_metastore/tables.py

@@ -116,24 +116,30 @@ def _crawl(self, catalog: str, database: str) -> list[Table]:
         tasks = []
         for _, table, _is_tmp in self._fetch(f"SHOW TABLES FROM {catalog}.{database}"):
             tasks.append(partial(self._describe, catalog, database, table))
-        return ThreadedExecution.gather(f"listing tables in {catalog}.{database}", tasks)
+        results = ThreadedExecution.gather(f"listing tables in {catalog}.{database}", tasks)
 
-    def _describe(self, catalog: str, database: str, table: str) -> Table:
+        return [x for x in results if x is not None]
+
+    def _describe(self, catalog: str, database: str, table: str) -> Table | None:
         """Fetches metadata like table type, data format, external table location,
         and the text of a view if specified for a specific table within the given
         catalog and database.
         """
-        describe = {}
         full_name = f"{catalog}.{database}.{table}"
-        logger.debug(f"[{full_name}] fetching table metadata")
-        for key, value, _ in self._fetch(f"DESCRIBE TABLE EXTENDED {full_name}"):
-            describe[key] = value
-        return Table(
-            catalog=describe["Catalog"],
-            database=database,
-            name=table,
-            object_type=describe["Type"],
-            table_format=describe.get("Provider", "").upper(),
-            location=describe.get("Location", None),
-            view_text=describe.get("View Text", None),
-        )
+        try:
+            logger.debug(f"[{full_name}] fetching table metadata")
+            describe = {}
+            for key, value, _ in self._fetch(f"DESCRIBE TABLE EXTENDED {full_name}"):
+                describe[key] = value
+            return Table(
+                catalog=describe["Catalog"],
+                database=database,
+                name=table,
+                object_type=describe["Type"],
+                table_format=describe.get("Provider", "").upper(),
+                location=describe.get("Location", None),
+                view_text=describe.get("View Text", None),
+            )
+        except RuntimeError as e:
+            logger.error(f"Couldn't fetch information for table {full_name} : {e}")
+            return None

src/databricks/labs/ucx/install.py

@@ -72,7 +72,7 @@ def __init__(self, ws: WorkspaceClient, *, prefix: str = "ucx", promtps: bool =
         self._dashboards = {}
 
     def run(self):
-        logger.info(f'Installing UCX v{__version__}')
+        logger.info(f"Installing UCX v{__version__}")
         self._configure()
         self._create_dashboards()
         self._create_jobs()
@@ -509,6 +509,6 @@ def _deployed_steps(self):
 
 if __name__ == "__main__":
     ws = WorkspaceClient(product="ucx", product_version=__version__)
-    logger.setLevel('INFO')
+    logger.setLevel("INFO")
     installer = Installer(ws)
     installer.run()

tests/unit/hive_metastore/test_tables.py

@@ -87,3 +87,13 @@ def test_uc_sql(table, query):
 def test_tables_crawler_inventory_table():
     tc = TablesCrawler(MockBackend(), "main", "default")
     assert tc._table == "tables"
+
+
+def test_tables_returning_error_when_describing():
+    errors = {"DESCRIBE TABLE EXTENDED test.database.table1": "error"}
+    rows = {
+        "SHOW TABLES FROM test.database": [("", "table1", ""), ("", "table2", "")],
+        "DESCRIBE TABLE EXTENDED test.database.table2": [("Catalog", "catalog", ""), ("Type", "delta", "")],
+    }
+    tc = TablesCrawler(MockBackend(fails_on_first=errors, rows=rows), "main", "default")
+    assert len(tc._crawl("test", "database")) == 1