def test_trims_seen_ids(self):
# If a relationship tries to follow to an ID we've already seen we
# should stop it
self.create_owner(1, 'Bob')
self.create_pet(1, 'Ginger', parent_id=None, owner_id=1)
relations = [
From('pet', 'owner_id').to('owner', 'id').bidirectional(),
]
# Mock out the get_table method so we can track how often it is called
original_get_table = dumper.Dumper._get_table
def mock_get_table(*args, **kwargs):
mock_get_table.call_count += 1
original_get_table(*args, **kwargs)
mock_get_table.call_count = 0
dumper.Dumper._get_table = mock_get_table
try:
self.do_partial_dump(relations, 'owner', '1=1')
# Only two tables, should only get called twice!
self.assertEquals(2, mock_get_table.call_count)
finally:
dumper.get_table = original_get_table
def test_lots_of_references(self):
# Lots of references should work fine
for x in xrange(1, 201):
self.create_owner(x, 'Bob')
self.create_pet(x, 'Ginger', parent_id=None, owner_id=x)
relations = [
From('pet', 'owner_id').to('owner', 'id').bidirectional(),
]
self.do_partial_dump(relations, 'owner', '1=1')
self.import_dump()
self.assertEquals(200, len(self.get_owners()))
self.assertEquals(200, len(self.get_pets()))
def test_unidirectional_link_forwards(self):
# Unidirectional links should only work one way. This is to account for
# tables not having indexes when following the link backwards and hence
# making the dump really slow
self.create_owner(1, 'Bob')
self.create_pet(1, 'Ginger', parent_id=None, owner_id=1)
relations = [
From('pet', 'owner_id').to('owner', 'id'),
]
self.do_partial_dump(relations, 'pet', '1=1')
self.import_dump()
self.assertEquals(1, len(self.get_owners()))
self.assertEquals(1, len(self.get_pets()))
def test_batch_size(self):
self.create_owner(1, 'Alan')
self.create_owner(2, 'Bob')
self.create_pet(1, 'Ginger', parent_id=None, owner_id=1)
self.create_pet(2, 'Tabby', parent_id=None, owner_id=2)
pks = {
'owner': Pk(['id']).in_batches(1),
'pet': Pk(['id']),
}
relations = [
From('owner', 'id').to('pet', 'owner_id'),
]
self.do_partial_dump(relations, 'owner', '1=1', pks=pks)
# Each owner should result in a distinct insert into the pet table
self.import_dump()
f = open("%s.%d" % (TEST_OUTPUT_PREFIX, 0), 'r')
result = f.read()
f.close()
self.assertEquals(1 + 2, result.count('INSERT'))
def test_many_rows_two_chunks(self):
# Creating two chunks and importing them should work fine
for x in xrange(1, 201):
self.create_owner(x, 'Bob')
self.create_pet(x, 'Ginger', parent_id=None, owner_id=x)
relations = [
From('pet', 'owner_id').to('owner', 'id').bidirectional(),
]
self.do_partial_dump(relations, 'owner', '1=1', chunks=2)
self.import_dump(chunks=2)
self.assertEquals(200, len(self.get_owners()))
self.assertEquals(200, len(self.get_pets()))
# The two files should be fairly sizable
size1 = os.path.getsize("%s.%d" % (TEST_OUTPUT_PREFIX, 0))
size2 = os.path.getsize("%s.%d" % (TEST_OUTPUT_PREFIX, 0))
self.assertTrue(size1 > 1000)
self.assertTrue(size2 > 1000)
def test_back_reference(self):
# A reference from X to Y should cause X be pulled in if Y is pulled in
self.create_owner(1, 'Bob')
self.create_pet(1, 'Ginger', parent_id=None, owner_id=1)
relations = [
From('pet', 'owner_id').to('owner', 'id').bidirectional(),
]
self.do_partial_dump(relations, 'owner', '1=1')
# Reimporting the result should give a single row that is the same as
# the original input
self.import_dump()
owners = self.get_owners()
self.assertEquals(1, len(owners))
pets = self.get_pets()
self.assertEquals(1, len(pets))
self.assertEquals('Ginger', pets[1]['name'])
self.assertEquals(None, pets[1]['parent_id'])
self.assertEquals(1, pets[1]['owner_id'])
def test_multiple_value_reference(self):
# A reference that pulls in multiple rows should work
self.create_owner(1, 'Bob')
self.create_pet(1, 'Ginger', parent_id=None, owner_id=1)
self.create_pet(2, 'Tabby', parent_id=None, owner_id=1)
relations = [
From('owner', 'id').to('pet', 'owner_id'),
]
self.do_partial_dump(relations, 'owner', '1=1')
# Reimporting the result should give a single row that is the same as
# the original input
self.import_dump()
pets = self.get_pets()
self.assertEquals(2, len(pets))
self.assertEquals('Ginger', pets[1]['name'])
self.assertEquals(None, pets[1]['parent_id'])
self.assertEquals(1, pets[1]['owner_id'])
self.assertEquals('Tabby', pets[2]['name'])
self.assertEquals(None, pets[2]['parent_id'])
self.assertEquals(1, pets[2]['owner_id'])
from mysqlpartialdump import Pk, From, NO_KEY_CACHE, ALLOW_DUPLICATES
def clean_email(row):
row['email'] = "%s%d" % (row['email'][:3], hash(row['email']))
return row
pks = {
'Customer': Pk(['id'], NO_KEY_CACHE, ALLOW_DUPLICATES),
'Order': Pk(['id']).in_batches(1),
'OrderLine': Pk(['id']),
'Product': Pk(['id']),
}
relationships = [
From('Customer', 'id').to('Order', 'customer_id').bidirectional(),
From('Order', 'id').to('OrderLine', 'order_id').bidirectional(),
From('OrderLine', 'product_id').to('Product', 'id').bidirectional(),
]
callbacks = {
'Customer': clean_email,
}
end_sql = ""
start_table = 'Product'
start_where = '1=1'
start_args = []
from mysqlpartialdump import Pk, From
pks = {
'Customer': Pk(['id']),
'Order': Pk(['id']),
'OrderLine': Pk(['id']),
'Product': Pk(['id']),
}
relationships = [
From('Customer', 'id').to('Order', 'customer_id'),
From('Order', 'id').to('OrderLine', 'order_id'),
From('OrderLine', 'product_id').to('Product', 'id'),
]
callbacks = {}
end_sql = ""
start_table = 'Customer'
start_where = 'id=%s'
start_args = ['1']