DH-4537/sorting the tables and column (#166)

dataherald/db/__init__.py

@@ -29,7 +29,7 @@ def find_by_id(self, collection: str, id: str) -> dict:
         pass
 
     @abstractmethod
-    def find(self, collection: str, query: dict) -> list:
+    def find(self, collection: str, query: dict, sort: list = None) -> list:
         pass
 
     @abstractmethod

dataherald/db/mongo.py

@@ -40,7 +40,9 @@ def find_by_id(self, collection: str, id: str) -> dict:
         return self._data_store[collection].find_one({"_id": ObjectId(id)})
 
     @override
-    def find(self, collection: str, query: dict) -> list:
+    def find(self, collection: str, query: dict, sort: list = None) -> list:
+        if sort:
+            return self._data_store[collection].find(query).sort(sort)
         return self._data_store[collection].find(query)
 
     @override

dataherald/db_scanner/repository/base.py

@@ -1,6 +1,7 @@
 from typing import List
 
 from bson.objectid import ObjectId
+from pymongo import ASCENDING
 
 from dataherald.db_scanner.models.types import TableSchemaDetail
 
@@ -68,10 +69,11 @@ def find_all(self) -> list[TableSchemaDetail]:
 
     def find_by(self, query: dict) -> list[TableSchemaDetail]:
         query = {k: v for k, v in query.items() if v}
-        rows = self.storage.find(DB_COLLECTION, query)
+        rows = self.storage.find(DB_COLLECTION, query, sort=[("table_name", ASCENDING)])
         result = []
         for row in rows:
             obj = TableSchemaDetail(**row)
+            obj.columns = sorted(obj.columns, key=lambda x: x.name)
             obj.id = str(row["_id"])
             result.append(obj)
         return result

dataherald/tests/db/test_db.py

@@ -48,7 +48,7 @@ def find_by_id(self, collection: str, id: str) -> dict:
         return None
 
     @override
-    def find(self, collection: str, query: dict) -> list:
+    def find(self, collection: str, query: dict, sort: list = None) -> list:
         return []
 
     @override