Better primary key

test_load_table.py

@@ -51,23 +51,23 @@ def test_good_csv(self, tmpdir):
         assert 'people' in self.meta.tables
         people = self.meta.tables['people']
 
-        assert_column_type(people.columns['index'], int)
+        assert_column_type(people.columns['id'], int)
         assert_column_type(people.columns['name'], str)
         assert_column_type(people.columns['dob'], str)
         assert_column_type(people.columns['number_of_pets'], str)
 
-        assert people.columns['index'].primary_key is True
+        assert people.columns['id'].primary_key is True
         assert people.columns['name'].primary_key is False
 
         assert_columns_have_non_unique_indexes(self.engine, 'people', 'name',
                                                'dob', 'number_of_pets')
 
         connection = self.engine.connect()
         results = connection.execute(
-            select([people]).order_by(people.c.index)).fetchall()
-        assert (0, 'Tom', '1980-02-26', '0') == results[0]
-        assert (1, 'Dick', '1982-03-14', '3') == results[1]
-        assert (2, 'Harry', '1972-11-24', '2') == results[2]
+            select([people]).order_by(people.c.id)).fetchall()
+        assert (1, 'Tom', '1980-02-26', '0') == results[0]
+        assert (2, 'Dick', '1982-03-14', '3') == results[1]
+        assert (3, 'Harry', '1972-11-24', '2') == results[2]
 
     def test_load_with_chunking(self, tmpdir):
         CSV = """name, dob, number_of_pets
@@ -82,10 +82,10 @@ def test_load_with_chunking(self, tmpdir):
 
         connection = self.engine.connect()
         results = connection.execute(
-            select([people]).order_by(people.c.index)).fetchall()
-        assert (0, 'Tom', '1980-02-26', '0') == results[0]
-        assert (1, 'Dick', '1982-03-14', '3') == results[1]
-        assert (2, 'Harry', '1972-11-24', '2') == results[2]
+            select([people]).order_by(people.c.id)).fetchall()
+        assert (1, 'Tom', '1980-02-26', '0') == results[0]
+        assert (2, 'Dick', '1982-03-14', '3') == results[1]
+        assert (3, 'Harry', '1972-11-24', '2') == results[2]
 
     def test_empty_csv(self, tmpdir):
         file_name = 'empty.csv'
@@ -120,10 +120,10 @@ def test_null_empty_string_handling(self, tmpdir):
 
         connection = self.engine.connect()
         results = connection.execute(
-            select([people]).order_by(people.c.index)).fetchall()
-        assert (0, 'Tom', '0') == results[0]
-        assert (1, None, '3') == results[1]
-        assert (2, None, '2') == results[2]
+            select([people]).order_by(people.c.id)).fetchall()
+        assert (1, 'Tom', '0') == results[0]
+        assert (2, None, '3') == results[1]
+        assert (3, None, '2') == results[2]
 
     def test_embedded_space_in_tablename_and_filename(self, tmpdir):
         CSV = """name, number_of_pets

utils.py

@@ -95,12 +95,13 @@ def load_table(filename,
                          dtype=dtypes,
                          skipinitialspace=True)
     for idx, chunk in enumerate(chunks):
-        chunk.index += chunksize * idx
+        chunk.index += chunksize * idx + 1
+        chunk.index.name = 'id'
         to_sql(tablename,
                engine,
                chunk,
                chunksize=chunksize,
-               keys='index',
+               keys='id',
                if_exists='append', )
     _index_table(tablename, metadata, engine, config.CASE_INSENSITIVE)
 
@@ -111,7 +112,7 @@ def _index_table(tablename, metadata, engine, case_insensitive=False):
     """
     table = sa.Table(tablename, metadata, autoload_with=engine)
     for label, column in table.columns.items():
-        if label == 'index':
+        if label == 'id':
             continue
         index_name = 'ix_{0}'.format(label.lower())
         indexes = [sa.Index(index_name, column)]