def test_expression_column_names(self):
expr = Expression('a.a', '==', 'b.b')
expected_column_names = set([ColumnName('a.a'), ColumnName('b.b')])
self.assertEqual(expected_column_names, expr.column_names)
expr = Expression('a.a', '==', '1')
expected_column_names = set([ColumnName('a.a')])
self.assertEqual(expected_column_names, expr.column_names)
def test_condition_column_names(self):
boolean_condition = AndList([
Expression('a.a', '==', 'b.b'),
Expression('a.a', '==', '1'),
Expression('1', '==', 'c.c'),
Expression('d.d', '==', '1'),
])
expected_column_names = set([
ColumnName('a.a'), ColumnName('b.b'), ColumnName('c.c'), ColumnName('d.d')
])
self.assertEqual(expected_column_names, boolean_condition.column_names)
def test_subset_rows(self):
conditions = [
[Expression('col_b', '==', '1'), 'or', Expression('col_a', '==', '2')]
]
self.table_a.subset_rows(conditions)
cmds_actual = self.table_a.cmds
cmds_expected = [
'echo -e "1,1\n2,3\n3,2"',
"awk -F',' 'OFS=\",\" { if (($2 == 1 || $1 == 2)) { print $1,$2 } }'"]
self.assertEqual(cmds_actual, cmds_expected)
def test_classify_conditions(self):
aliases = {'a': 'a', 'alpha': 'a', 'b': 'b', 'c': 'c'}
conditions = [
{
'left_operand': 'a.col1',
'operator': '=',
'right_operand': 'b.col1'
},
'and',
{
'left_operand': 'a.col1',
'operator': '=',
'right_operand': 5
},
'and',
{
'left_operand': 'a.col1',
'operator': '=',
'right_operand': 'c.col1'
},
'and',
{
'left_operand': 'a.col1',
'operator': '=',
'right_operand': 'a.col2'
},
]
expected_join_conditions = [
Expression(ColumnName('a.col1'), '=', ColumnName('b.col1')),
Expression(ColumnName('a.col1'), '=', ColumnName('c.col1')),
]
expected_where_conditions = [
Expression(ColumnName('a.col1'), '=', 5),
Expression(ColumnName('a.col1'), '=', ColumnName('a.col2')),
]
actual_join_conditions, actual_where_conditions = classify_conditions(
conditions)
self.assertEqual(actual_join_conditions, expected_join_conditions)
self.assertEqual(actual_where_conditions, expected_where_conditions)
def test_build_graph(self):
relations = [
{
'path': 'a.txt',
'alias': 'a'
},
{
'path': 'b.txt',
'alias': 'b'
},
{
'path': 'c.txt',
'alias': 'c'
},
]
tables = [Table(r['path'], alias=r['alias']) for r in relations]
aliases = {'a': 'a', 'alpha': 'a', 'b': 'b', 'c': 'c'}
conditions = [
Expression(ColumnName('a.col1'), '=', ColumnName('b.col1')),
Expression(ColumnName('a.col1'), '=', ColumnName('c.col1')),
]
expected_graph = {
'a': {
'idx': 0,
'neighbors': set(['b', 'c']),
},
'b': {
'idx': 1,
'neighbors': set(['a']),
},
'c': {
'idx': 2,
'neighbors': set(['a']),
},
}
actual_graph = build_graph(tables, conditions)
self.assertEqual(actual_graph, expected_graph)
def test_stage_conditions(self):
conditions = [
Expression('table_a.col_a', '==', 'table_b.col_z'),
Expression('table_a.col_a', '==', 'table_a.col_b'),
OrList([
Expression('table_a.col_a', '==', 'table_b.col_z'),
Expression('table_a.col_a', '==', 'table_a.col_b')
]),
]
expected_condition_order = [
[Expression('table_a.col_a', '==', 'table_a.col_b')],
[
Expression('table_a.col_a', '==', 'table_b.col_z'),
OrList([
Expression('table_a.col_a', '==', 'table_b.col_z'),
Expression('table_a.col_a', '==', 'table_a.col_b')
])
],
]
actual_condition_order = stage_conditions([self.table_a, self.table_b], conditions)
self.assertEqual(expected_condition_order, actual_condition_order)
def test_condition_applies(self):
condition = AndList([
Expression('table_a.col_a', '==', 'table_b.col_z'),
Expression('table_a.col_b', '==', 'table_b.col_a')
])
self.assertTrue(condition_applies(condition, self.table_a, self.table_b))
condition = OrList([
Expression('table_a.col_a', '==', 'table_b.col_z'),
Expression('table_a.col_b', '==', 'table_b.col_a')
])
self.assertTrue(condition_applies(condition, self.table_a, self.table_b))
condition = OrList([
Expression('table_a.col_a', '==', 'table_b.col_z'),
Expression('table_a.col_b', '==', 'table_b.col_a')
])
self.assertFalse(condition_applies(condition, self.table_a))
condition = AndList([
Expression('table_a.col_a', '==', 'table_b.col_z'),
Expression('table_c.col_b', '==', 'table_b.col_a')
])
self.assertFalse(condition_applies(condition, self.table_a, self.table_b))
condition = OrList([
Expression('table_a.col_a', '==', 'table_b.col_j'),
Expression('table_a.col_b', '==', 'table_b.col_a')
])
self.assertFalse(condition_applies(condition, self.table_a, self.table_b))
condition = AndList([
Expression('table_a.col_a', '==', '1'),
Expression('table_a.col_b', '==', 'table_b.col_a')
])
self.assertTrue(condition_applies(condition, self.table_a, self.table_b))
with self.assertRaises(AmbiguousColumnNameError):
condition = AndList([
Expression('table_a.col_a', '==', '1'),
Expression('table_a.col_b', '==', 'col_a')
])
condition_applies(condition, self.table_a, self.table_b)