Merge pull request #1797 from dbcli/RW/add-parseutils-tests-02

Increase test coverage for `parseutils.py`

Author: rolandwalker

test/pytests/test_parseutils.py

@@ -2,7 +2,10 @@
 
 import pytest
 import sqlparse
+from sqlparse.sql import Identifier, IdentifierList, Token, TokenList
+from sqlparse.tokens import DML, Keyword, Punctuation
 
+from mycli.packages import parseutils
 from mycli.packages.parseutils import (
     extract_columns_from_select,
     extract_from_part,
@@ -280,6 +283,18 @@ def test_extract_from_part_handles_multiple_joins_and_skips_on_clause():
     assert token_values(tokens) == ['abc', 'join', 'def', 'ghi']
 
 
+def test_extract_from_part_recurses_into_subselect_and_stops_at_punctuation():
+    parsed = sqlparse.parse('select * from (select * from inner_table), outer_table')[0]
+    tokens = extract_from_part(parsed)
+    assert token_values(tokens) == ['inner_table']
+
+
+def test_extract_from_part_stops_at_punctuation_when_requested():
+    parsed = TokenList([Token(Keyword, 'FROM'), Token(Punctuation, ','), Token(Keyword, 'SELECT')])
+    tokens = extract_from_part(parsed, stop_at_punctuation=True)
+    assert token_values(tokens) == []
+
+
 def test_extract_table_identifiers_handles_identifier_list():
     parsed = sqlparse.parse('select * from abc a, def d')[0]
     token_stream = extract_from_part(parsed)
@@ -301,6 +316,33 @@ def test_extract_table_identifiers_handles_function_tokens():
     assert list(extract_table_identifiers(token_stream)) == [(None, 'my_func', 'my_func')]
 
 
+def test_extract_table_identifiers_skips_identifier_list_entries_without_identifier_methods():
+    class BrokenIdentifierList(IdentifierList):
+        def get_identifiers(self):
+            return [object()]
+
+    assert list(extract_table_identifiers(iter([BrokenIdentifierList([])]))) == []
+
+
+def test_extract_table_identifiers_uses_name_when_identifier_has_no_real_name():
+    class NamelessIdentifier(Identifier):
+        def get_real_name(self):
+            return None
+
+        def get_parent_name(self):
+            return None
+
+        def get_name(self):
+            return 'fallback_name'
+
+        def get_alias(self):
+            return None
+
+    assert list(extract_table_identifiers(iter([NamelessIdentifier([])]))) == [
+        (None, 'fallback_name', 'fallback_name'),
+    ]
+
+
 @pytest.mark.parametrize(
     ('sql', 'expected_keyword', 'expected_text'),
     [
@@ -335,6 +377,82 @@ def test_query_is_single_table_update(sql, is_single_table):
     assert query_is_single_table_update(sql) is is_single_table
 
 
+def test_extract_columns_from_select_handles_falsey_last_select(monkeypatch):
+    monkeypatch.setattr(parseutils, 'get_last_select', lambda _parsed: [])
+    assert extract_columns_from_select('select 1') == []
+
+
+def test_extract_columns_from_select_handles_single_identifier(monkeypatch):
+    class SingleIdentifier(Identifier):
+        def get_real_name(self):
+            return 'column_name'
+
+    monkeypatch.setattr(
+        parseutils,
+        'get_last_select',
+        lambda _parsed: TokenList([Token(DML, 'SELECT'), SingleIdentifier([])]),
+    )
+
+    assert extract_columns_from_select('select column_name') == ['column_name']
+
+
+def test_extract_columns_from_select_ignores_unhandled_identifier_list_entries(monkeypatch):
+    class WeirdIdentifierList(IdentifierList):
+        def get_identifiers(self):
+            return [object()]
+
+    monkeypatch.setattr(
+        parseutils,
+        'get_last_select',
+        lambda _parsed: TokenList([Token(DML, 'SELECT'), WeirdIdentifierList([])]),
+    )
+
+    assert extract_columns_from_select('select 1') == []
+
+
+def test_extract_columns_from_select_stops_at_keyword_before_collecting_columns(monkeypatch):
+    monkeypatch.setattr(
+        parseutils,
+        'get_last_select',
+        lambda _parsed: TokenList([Token(DML, 'SELECT'), Token(Keyword, 'FROM')]),
+    )
+
+    assert extract_columns_from_select('select 1') == []
+
+
+def test_extract_tables_from_complete_statements_returns_empty_for_falsey_rough_parse(monkeypatch):
+    monkeypatch.setattr(parseutils.sqlparse, 'parse', lambda _sql: [])
+
+    assert extract_tables_from_complete_statements('select * from t') == []
+
+
+def test_extract_tables_from_complete_statements_skips_cte_table_identifiers(monkeypatch):
+    class FakeParentSelect:
+        def sql(self):
+            return 'WITH cte AS (SELECT 1) SELECT * FROM cte'
+
+    class FakeIdentifier:
+        parent_select = FakeParentSelect()
+        db = ''
+        name = 'cte'
+        alias = ''
+
+    class FakeStatement:
+        def find_all(self, _table_type):
+            return [FakeIdentifier()]
+
+    monkeypatch.setattr(parseutils.sqlparse, 'parse', lambda _sql: ['stmt'])
+    monkeypatch.setattr(parseutils.sqlglot, 'parse_one', lambda *_args, **_kwargs: FakeStatement())
+
+    assert extract_tables_from_complete_statements('with cte as (select 1) select * from cte') == []
+
+
+def test_query_is_single_table_update_returns_false_when_parse_result_is_empty(monkeypatch):
+    monkeypatch.setattr(parseutils.sqlparse, 'parse', lambda _sql: [])
+
+    assert query_is_single_table_update('update test set x = 1') is False
+
+
 def test_is_destructive():
     sql = "use test;\nshow databases;\ndrop database foo;"
     assert is_destructive(["drop"], sql) is True
@@ -360,6 +478,16 @@ def test_is_destructive_update_without_where_clause():
     assert is_destructive(["update"], sql) is True
 
 
+def test_is_destructive_skips_empty_split_queries(monkeypatch):
+    monkeypatch.setattr(parseutils.sqlparse, 'split', lambda _queries: ['', ''])
+
+    assert is_destructive(['drop'], 'ignored') is False
+
+
+def test_is_destructive_returns_false_when_no_query_matches_keywords() -> None:
+    assert is_destructive(['drop'], 'select 1; show databases;') is False
+
+
 @pytest.mark.parametrize(
     ("sql", "has_where_clause"),
     [
@@ -389,3 +517,7 @@ def test_query_has_where_clause(sql, has_where_clause):
 )
 def test_is_dropping_database(sql, dbname, is_dropping):
     assert is_dropping_database(sql, dbname) == is_dropping
+
+
+def test_is_dropping_database_skips_statements_without_enough_keywords():
+    assert is_dropping_database('drop foo', 'foo') is False