Files
mongo/test/suite/test_bulk01.py
2016-01-01 16:37:39 -05:00

220 lines
8.9 KiB
Python

#!/usr/bin/env python
#
# Public Domain 2014-2016 MongoDB, Inc.
# Public Domain 2008-2014 WiredTiger, Inc.
#
# This is free and unencumbered software released into the public domain.
#
# Anyone is free to copy, modify, publish, use, compile, sell, or
# distribute this software, either in source code form or as a compiled
# binary, for any purpose, commercial or non-commercial, and by any
# means.
#
# In jurisdictions that recognize copyright laws, the author or authors
# of this software dedicate any and all copyright interest in the
# software to the public domain. We make this dedication for the benefit
# of the public at large and to the detriment of our heirs and
# successors. We intend this dedication to be an overt act of
# relinquishment in perpetuity of all present and future rights to this
# software under copyright law.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
# EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
# IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR
# OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
# ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
# OTHER DEALINGS IN THE SOFTWARE.
#
# test_bulk.py
# bulk-cursor test.
#
import wiredtiger, wttest
from helper import key_populate, value_populate
from wtscenario import check_scenarios, multiply_scenarios, number_scenarios
# Smoke test bulk-load.
class test_bulk_load(wttest.WiredTigerTestCase):
name = 'test_bulk'
types = [
('file', dict(type='file:')),
('table', dict(type='table:'))
]
keyfmt = [
('integer', dict(keyfmt='i')),
('recno', dict(keyfmt='r')),
('string', dict(keyfmt='S')),
]
valfmt = [
('fixed', dict(valfmt='8t')),
('integer', dict(valfmt='i')),
('string', dict(valfmt='S')),
]
scenarios = number_scenarios(multiply_scenarios('.', types, keyfmt, valfmt))
# Test a simple bulk-load
def test_bulk_load(self):
uri = self.type + self.name
self.session.create(uri,
'key_format=' + self.keyfmt + ',value_format=' + self.valfmt)
cursor = self.session.open_cursor(uri, None, "bulk")
for i in range(1, 1000):
cursor[key_populate(cursor, i)] = value_populate(cursor, i)
# Test a bulk-load triggers variable-length column-store RLE correctly.
def test_bulk_load_var_rle(self):
if self.keyfmt != 'r' or self.valfmt == '8t':
return
# We can't directly test RLE, it's internal to WiredTiger. However,
# diagnostic builds catch records that should have been RLE compressed,
# but weren't, so setting matching values should be sufficient.
uri = self.type + self.name
self.session.create(uri,
'key_format=' + self.keyfmt + ',value_format=' + self.valfmt)
cursor = self.session.open_cursor(uri, None, "bulk")
for i in range(1, 1000):
cursor[key_populate(cursor, i)] = value_populate(cursor, i/7)
# Test a bulk-load variable-length column-store append ignores any key.
def test_bulk_load_var_append(self):
if self.keyfmt != 'r':
return
uri = self.type + self.name
self.session.create(uri,
'key_format=' + self.keyfmt + ',value_format=' + self.valfmt)
cursor = self.session.open_cursor(uri, None, "bulk,append")
for i in range(1, 1000):
cursor[key_populate(cursor, 37)] = value_populate(cursor, i)
cursor.close()
cursor = self.session.open_cursor(uri, None, None)
for i in range(1, 1000):
cursor.set_key(key_populate(cursor, i))
cursor.search()
self.assertEqual(cursor.get_value(), value_populate(cursor, i))
# Test that column-store bulk-load handles skipped records correctly.
def test_bulk_load_col_delete(self):
if self.keyfmt != 'r':
return
uri = self.type + self.name
self.session.create(uri,
'key_format=' + self.keyfmt + ',value_format=' + self.valfmt)
cursor = self.session.open_cursor(uri, None, "bulk")
for i in range(1, 1000):
if i % 7 == 0:
cursor[key_populate(cursor, i)] = value_populate(cursor, i)
# Ensure we create all the missing records.
i = i + 1
cursor[key_populate(cursor, i)] = value_populate(cursor, i)
cursor.close()
cursor = self.session.open_cursor(uri, None, None)
# Verify all the records are there, in their proper state.
for i in range(1, 1000):
cursor.set_key(key_populate(cursor, i))
if i % 7 == 0:
cursor.search()
self.assertEqual(cursor.get_value(), value_populate(cursor, i))
elif cursor.value_format == '8t':
cursor.search()
self.assertEqual(cursor.get_value(), 0)
else:
self.assertEqual(cursor.search(), wiredtiger.WT_NOTFOUND)
# Test that variable-length column-store bulk-load efficiently creates big
# records.
def test_bulk_load_col_big(self):
if self.keyfmt != 'r' or self.valfmt == '8t':
return
uri = self.type + self.name
self.session.create(uri,
'key_format=' + self.keyfmt + ',value_format=' + self.valfmt)
cursor = self.session.open_cursor(uri, None, "bulk")
for i in range(1, 10):
cursor[key_populate(cursor, i)] = value_populate(cursor, i)
# A big record -- if it's not efficient, we'll just hang.
big = 18446744073709551606
cursor[key_populate(cursor, big)] = value_populate(cursor, big)
cursor.close()
cursor = self.session.open_cursor(uri, None, None)
cursor.set_key(key_populate(cursor, big))
cursor.search()
self.assertEqual(cursor.get_value(), value_populate(cursor, big))
# Test that bulk-load out-of-order fails.
def test_bulk_load_order_check(self):
uri = self.type + self.name
self.session.create(uri,
'key_format=' + self.keyfmt + ',value_format=' + self.valfmt)
cursor = self.session.open_cursor(uri, None, "bulk")
cursor[key_populate(cursor, 10)] = value_populate(cursor, 10)
for i in [1, 9, 10]:
cursor.set_key(key_populate(cursor, 1))
cursor.set_value(value_populate(cursor, 1))
msg = '/than previously inserted key/'
self.assertRaisesWithMessage(
wiredtiger.WiredTigerError, lambda: cursor.insert(), msg)
cursor[key_populate(cursor, 11)] = value_populate(cursor, 11)
# Test that row-store bulk-load out-of-order can succeed.
def test_bulk_load_row_order_nocheck(self):
# Row-store offers an optional fast-past that skips the relatively
# expensive key-order checks, used when the input is known to be
# correct. Column-store comparisons are cheap, so it doesn't have
# that fast-path support.
if self.keyfmt != 'S':
return
uri = self.type + self.name
self.session.create(uri,
'key_format=' + self.keyfmt + ',value_format=' + self.valfmt)
cursor = self.session.open_cursor(uri, None, "bulk,skip_sort_check")
cursor[key_populate(cursor, 10)] = value_populate(cursor, 10)
cursor[key_populate(cursor, 1)] = value_populate(cursor, 1)
if not wiredtiger.diagnostic_build():
self.skipTest('requires a diagnostic build')
# Close explicitly, there's going to be a fallure.
msg = '/are incorrectly sorted/'
self.assertRaisesWithMessage(wiredtiger.WiredTigerError,
lambda: self.conn.close(), msg)
# Test bulk-load only permitted on newly created objects.
def test_bulk_load_not_empty(self):
uri = self.type + self.name
self.session.create(uri, 'key_format=S,value_format=S')
cursor = self.session.open_cursor(uri, None)
cursor[key_populate(cursor, 1)] = value_populate(cursor, 1)
# Close the insert cursor, else we'll get EBUSY.
cursor.close()
msg = '/bulk-load is only supported on newly created objects/'
self.assertRaisesWithMessage(wiredtiger.WiredTigerError,
lambda: self.session.open_cursor(uri, None, "bulk"), msg)
# Test that bulk-load objects cannot be opened by other cursors.
def test_bulk_load_busy(self):
uri = self.type + self.name
self.session.create(uri, 'key_format=S,value_format=S')
cursor = self.session.open_cursor(uri, None)
cursor[key_populate(cursor, 1)] = value_populate(cursor, 1)
# Don't close the insert cursor, we want EBUSY.
self.assertRaises(wiredtiger.WiredTigerError,
lambda: self.session.open_cursor(uri, None, "bulk"))
if __name__ == '__main__':
wttest.run()