class testQueryTimeout(FlowTestsBase):
def __init__(self):
# skip test if we're running under Valgrind
if Env().envRunner.debugger is not None:
Env().skip() # queries will be much slower under Valgrind
self.env = Env(decodeResponses=True)
global redis_con
global redis_graph
redis_con = self.env.getConnection()
redis_graph = Graph("timeout", redis_con)
def test01_read_query_timeout(self):
query = "UNWIND range(0,100000) AS x WITH x AS x WHERE x = 10000 RETURN x"
try:
# The query is expected to time out
redis_graph.query(query, timeout=1)
assert (False)
except ResponseError as error:
self.env.assertContains("Query timed out", str(error))
try:
# The query is expected to succeed
redis_graph.query(query, timeout=100)
except:
assert (False)
def test02_configured_timeout(self):
# Verify that the module-level timeout is set to the default of 0
response = redis_con.execute_command("GRAPH.CONFIG GET timeout")
self.env.assertEquals(response[1], 0)
# Set a default timeout of 1 millisecond
redis_con.execute_command("GRAPH.CONFIG SET timeout 1")
response = redis_con.execute_command("GRAPH.CONFIG GET timeout")
self.env.assertEquals(response[1], 1)
# Validate that a read query times out
query = "UNWIND range(0,100000) AS x WITH x AS x WHERE x = 10000 RETURN x"
try:
redis_graph.query(query)
assert (False)
except ResponseError as error:
self.env.assertContains("Query timed out", str(error))
def test03_write_query_ignore_timeout(self):
# Verify that the timeout argument is ignored by write queries
query = "CREATE (a:M) WITH a UNWIND range(1,10000) AS ctr SET a.v = ctr"
try:
# The query should complete successfully
actual_result = redis_graph.query(query, timeout=1)
# The query should have taken longer than the timeout value
self.env.assertGreater(actual_result.run_time_ms, 1)
# The query should have updated properties 10,000 times
self.env.assertEquals(actual_result.properties_set, 10000)
except ResponseError:
assert (False)
class TestAggregate():
def __init__(self):
self.env = Env()
add_values(self.env)
def testGroupBy(self):
cmd = [
'ft.aggregate', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'count', '0', 'AS', 'count', 'SORTBY', 2, '@count', 'desc',
'LIMIT', '0', '5'
]
res = self.env.cmd(*cmd)
self.env.assertIsNotNone(res)
self.env.assertEqual([
292L, ['brand', '', 'count', '1518'],
['brand', 'mad catz', 'count', '43'],
['brand', 'generic', 'count', '40'],
['brand', 'steelseries', 'count', '37'],
['brand', 'logitech', 'count', '35']
], res)
def testMinMax(self):
cmd = [
'ft.aggregate', 'games', 'sony', 'GROUPBY', '1', '@brand',
'REDUCE', 'count', '0', 'REDUCE', 'min', '1', '@price', 'as',
'minPrice', 'SORTBY', '2', '@minPrice', 'DESC'
]
res = self.env.cmd(*cmd)
self.env.assertIsNotNone(res)
row = to_dict(res[1])
self.env.assertEqual(88, int(float(row['minPrice'])))
cmd = [
'ft.aggregate', 'games', 'sony', 'GROUPBY', '1', '@brand',
'REDUCE', 'count', '0', 'REDUCE', 'max', '1', '@price', 'as',
'maxPrice', 'SORTBY', '2', '@maxPrice', 'DESC'
]
res = self.env.cmd(*cmd)
row = to_dict(res[1])
self.env.assertEqual(695, int(float(row['maxPrice'])))
def testAvg(self):
cmd = [
'ft.aggregate', 'games', 'sony', 'GROUPBY', '1', '@brand',
'REDUCE', 'avg', '1', '@price', 'AS', 'avg_price', 'REDUCE',
'count', '0', 'SORTBY', '2', '@avg_price', 'DESC'
]
res = self.env.cmd(*cmd)
self.env.assertIsNotNone(res)
self.env.assertEqual(26, res[0])
# Ensure the formatting actually exists
first_row = to_dict(res[1])
self.env.assertEqual(109, int(float(first_row['avg_price'])))
for row in res[1:]:
row = to_dict(row)
self.env.assertIn('avg_price', row)
# Test aliasing
cmd = [
'FT.AGGREGATE', 'games', 'sony', 'GROUPBY', '1', '@brand',
'REDUCE', 'avg', '1', '@price', 'AS', 'avgPrice'
]
res = self.env.cmd(*cmd)
first_row = to_dict(res[1])
self.env.assertEqual(17, int(float(first_row['avgPrice'])))
def testCountDistinct(self):
cmd = [
'FT.AGGREGATE', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'COUNT_DISTINCT', '1', '@title', 'AS', 'count_distinct(title)',
'REDUCE', 'COUNT', '0'
]
res = self.env.cmd(*cmd)[1:]
# print res
row = to_dict(res[0])
self.env.assertEqual(1484, int(row['count_distinct(title)']))
cmd = [
'FT.AGGREGATE', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'COUNT_DISTINCTISH', '1', '@title', 'AS',
'count_distinctish(title)', 'REDUCE', 'COUNT', '0'
]
res = self.env.cmd(*cmd)[1:]
# print res
row = to_dict(res[0])
self.env.assertEqual(1461, int(row['count_distinctish(title)']))
def testQuantile(self):
cmd = [
'FT.AGGREGATE', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'QUANTILE', '2', '@price', '0.50', 'AS', 'q50', 'REDUCE',
'QUANTILE', '2', '@price', '0.90', 'AS', 'q90', 'REDUCE',
'QUANTILE', '2', '@price', '0.95', 'AS', 'q95', 'REDUCE', 'AVG',
'1', '@price', 'REDUCE', 'COUNT', '0', 'AS', 'rowcount', 'SORTBY',
'2', '@rowcount', 'DESC', 'MAX', '1'
]
res = self.env.cmd(*cmd)
row = to_dict(res[1])
# TODO: Better samples
self.env.assertAlmostEqual(14.99, float(row['q50']), delta=3)
self.env.assertAlmostEqual(70, float(row['q90']), delta=50)
# This tests the 95th percentile, which is error prone because
# so few samples actually exist. I'm disabling it for now so that
# there is no breakage in CI
# self.env.assertAlmostEqual(110, (float(row['q95'])), delta=50)
def testStdDev(self):
cmd = [
'FT.AGGREGATE', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'STDDEV', '1', '@price', 'AS', 'stddev(price)', 'REDUCE', 'AVG',
'1', '@price', 'AS', 'avgPrice', 'REDUCE', 'QUANTILE', '2',
'@price', '0.50', 'AS', 'q50Price', 'REDUCE', 'COUNT', '0', 'AS',
'rowcount', 'SORTBY', '2', '@rowcount', 'DESC', 'LIMIT', '0', '10'
]
res = self.env.cmd(*cmd)
row = to_dict(res[1])
self.env.assertTrue(10 <= int(float(row['q50Price'])) <= 20)
self.env.assertAlmostEqual(53,
int(float(row['stddev(price)'])),
delta=50)
self.env.assertEqual(29, int(float(row['avgPrice'])))
def testParseTime(self):
cmd = [
'FT.AGGREGATE', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'COUNT', '0', 'AS', 'count', 'APPLY', 'timefmt(1517417144)', 'AS',
'dt', 'APPLY', 'parse_time("%FT%TZ", @dt)', 'as', 'parsed_dt',
'LIMIT', '0', '1'
]
res = self.env.cmd(*cmd)
self.env.assertEqual([
'brand', '', 'count', '1518', 'dt', '2018-01-31T16:45:44Z',
'parsed_dt', '1517417144'
], res[1])
def testRandomSample(self):
cmd = [
'FT.AGGREGATE', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'COUNT', '0', 'AS', 'num', 'REDUCE', 'RANDOM_SAMPLE', '2',
'@price', '10', 'SORTBY', '2', '@num', 'DESC', 'MAX', '10'
]
for row in self.env.cmd(*cmd)[1:]:
self.env.assertIsInstance(row[5], list)
self.env.assertGreater(len(row[5]), 0)
self.env.assertGreaterEqual(row[3], len(row[5]))
self.env.assertLessEqual(len(row[5]), 10)
def testTimeFunctions(self):
cmd = [
'FT.AGGREGATE', 'games', '*', 'APPLY', '1517417144', 'AS', 'dt',
'APPLY', 'timefmt(@dt)', 'AS', 'timefmt', 'APPLY', 'day(@dt)',
'AS', 'day', 'APPLY', 'hour(@dt)', 'AS', 'hour', 'APPLY',
'minute(@dt)', 'AS', 'minute', 'APPLY', 'month(@dt)', 'AS',
'month', 'APPLY', 'dayofweek(@dt)', 'AS', 'dayofweek', 'APPLY',
'dayofmonth(@dt)', 'AS', 'dayofmonth', 'APPLY', 'dayofyear(@dt)',
'AS', 'dayofyear', 'APPLY', 'year(@dt)', 'AS', 'year', 'LIMIT',
'0', '1'
]
res = self.env.cmd(*cmd)
self.env.assertListEqual([
1L,
[
'dt', '1517417144', 'timefmt', '2018-01-31T16:45:44Z', 'day',
'1517356800', 'hour', '1517414400', 'minute', '1517417100',
'month', '1514764800', 'dayofweek', '3', 'dayofmonth', '31',
'dayofyear', '30', 'year', '2018'
]
], res)
def testStringFormat(self):
cmd = [
'FT.AGGREGATE', 'games', '@brand:sony', 'GROUPBY', '2', '@title',
'@brand', 'REDUCE', 'COUNT', '0', 'REDUCE', 'MAX', '1', '@price',
'AS', 'price', 'APPLY',
'format("%s|%s|%s|%s", @title, @brand, "Mark", @price)', 'as',
'titleBrand', 'LIMIT', '0', '10'
]
res = self.env.cmd(*cmd)
for row in res[1:]:
row = to_dict(row)
expected = '%s|%s|%s|%g' % (row['title'], row['brand'], 'Mark',
float(row['price']))
self.env.assertEqual(expected, row['titleBrand'])
def testSum(self):
cmd = [
'ft.aggregate', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'count', '0', 'AS', 'count', 'REDUCE', 'sum', 1, '@price', 'AS',
'sum(price)', 'SORTBY', 2, '@sum(price)', 'desc', 'LIMIT', '0', '5'
]
res = self.env.cmd(*cmd)
self.env.assertEqual([
292L, ['brand', '', 'count', '1518', 'sum(price)', '44780.69'],
['brand', 'mad catz', 'count', '43', 'sum(price)', '3973.48'],
['brand', 'razer', 'count', '26', 'sum(price)', '2558.58'],
['brand', 'logitech', 'count', '35', 'sum(price)', '2329.21'],
['brand', 'steelseries', 'count', '37', 'sum(price)', '1851.12']
], res)
def testFilter(self):
cmd = [
'ft.aggregate', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'count', '0', 'AS', 'count', 'FILTER', '@count > 5'
]
res = self.env.cmd(*cmd)
for row in res[1:]:
row = to_dict(row)
self.env.assertGreater(int(row['count']), 5)
cmd = [
'ft.aggregate', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'count', '0', 'AS', 'count', 'FILTER', '@count < 5', 'FILTER',
'@count > 2 && @brand != ""'
]
res = self.env.cmd(*cmd)
for row in res[1:]:
row = to_dict(row)
self.env.assertLess(int(row['count']), 5)
self.env.assertGreater(int(row['count']), 2)
def testToList(self):
cmd = [
'ft.aggregate', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'count_distinct', '1', '@price', 'as', 'count', 'REDUCE', 'tolist',
1, '@price', 'as', 'prices', 'SORTBY', 2, '@count', 'desc',
'LIMIT', '0', '5'
]
res = self.env.cmd(*cmd)
for row in res[1:]:
row = to_dict(row)
self.env.assertEqual(int(row['count']), len(row['prices']))
def testSortBy(self):
res = self.env.cmd('ft.aggregate', 'games', '*', 'GROUPBY', '1',
'@brand', 'REDUCE', 'sum', 1, '@price', 'as',
'price', 'SORTBY', 2, '@price', 'desc', 'LIMIT',
'0', '2')
self.env.assertListEqual([
292L, ['brand', '', 'price', '44780.69'],
['brand', 'mad catz', 'price', '3973.48']
], res)
res = self.env.cmd('ft.aggregate', 'games', '*', 'GROUPBY', '1',
'@brand', 'REDUCE', 'sum', 1, '@price', 'as',
'price', 'SORTBY', 2, '@price', 'asc', 'LIMIT', '0',
'2')
self.env.assertListEqual([
292L, ['brand', 'myiico', 'price', '0.23'],
['brand', 'crystal dynamics', 'price', '0.25']
], res)
# Test MAX with limit higher than it
res = self.env.cmd('ft.aggregate', 'games', '*', 'GROUPBY', '1',
'@brand', 'REDUCE', 'sum', 1, '@price', 'as',
'price', 'SORTBY', 2, '@price', 'asc', 'MAX', 2)
self.env.assertListEqual([
292L, ['brand', 'myiico', 'price', '0.23'],
['brand', 'crystal dynamics', 'price', '0.25']
], res)
# Test Sorting by multiple properties
res = self.env.cmd(
'ft.aggregate',
'games',
'*',
'GROUPBY',
'1',
'@brand',
'REDUCE',
'sum',
1,
'@price',
'as',
'price',
'APPLY',
'(@price % 10)',
'AS',
'price',
'SORTBY',
4,
'@price',
'asc',
'@brand',
'desc',
'MAX',
10,
)
self.env.assertListEqual([
292L, ['brand', 'zps', 'price', '0'],
['brand', 'zalman', 'price', '0'], [
'brand', 'yoozoo', 'price', '0'
], ['brand', 'white label', 'price', '0'],
['brand', 'stinky', 'price', '0'],
['brand', 'polaroid', 'price', '0'],
['brand', 'plantronics', 'price', '0'],
['brand', 'ozone', 'price', '0'], ['brand', 'oooo', 'price', '0'],
['brand', 'neon', 'price', '0']
], res)
def testExpressions(self):
pass
def testNoGroup(self):
res = self.env.cmd(
'ft.aggregate',
'games',
'*',
'LOAD',
'2',
'@brand',
'@price',
'APPLY',
'floor(sqrt(@price)) % 10',
'AS',
'price',
'SORTBY',
4,
'@price',
'desc',
'@brand',
'desc',
'MAX',
5,
)
exp = [
2265L, ['brand', 'xbox', 'price', '9'],
['brand', 'turtle beach', 'price', '9'],
['brand', 'trust', 'price', '9'],
['brand', 'steelseries', 'price', '9'],
['brand', 'speedlink', 'price', '9']
]
# exp = [2265L, ['brand', 'Xbox', 'price', '9'], ['brand', 'Turtle Beach', 'price', '9'], [
# 'brand', 'Trust', 'price', '9'], ['brand', 'SteelSeries', 'price', '9'], ['brand', 'Speedlink', 'price', '9']]
self.env.assertListEqual(exp[1], res[1])
def testLoad(self):
res = self.env.cmd('ft.aggregate', 'games', '*', 'LOAD', '3', '@brand',
'@price', '@nonexist', 'SORTBY', 2, '@price',
'DESC', 'MAX', 2)
exp = [
3L, ['brand', '', 'price', '759.12'],
['brand', 'Sony', 'price', '695.8']
]
self.env.assertEqual(exp[1], res[1])
def testSplit(self):
res = self.env.cmd(
'ft.aggregate', 'games', '*', 'APPLY',
'split("hello world, foo,,,bar,", ",", " ")', 'AS', 'strs',
'APPLY', 'split("hello world, foo,,,bar,", " ", ",")', 'AS',
'strs2', 'APPLY', 'split("hello world, foo,,,bar,", "", "")',
'AS', 'strs3', 'APPLY', 'split("hello world, foo,,,bar,")', 'AS',
'strs4', 'APPLY', 'split("hello world, foo,,,bar,",",")', 'AS',
'strs5', 'APPLY', 'split("")', 'AS', 'empty', 'LIMIT', '0', '1')
# print "Got {} results".format(len(res))
# return
# pprint.pprint(res)
self.env.assertListEqual([
1L,
[
'strs', ['hello world', 'foo', 'bar'], 'strs2',
['hello', 'world', 'foo,,,bar'], 'strs3',
['hello world, foo,,,bar,'], 'strs4',
['hello world', 'foo', 'bar'], 'strs5',
['hello world', 'foo', 'bar'], 'empty', []
]
], res)
def testFirstValue(self):
res = self.env.cmd(
'ft.aggregate', 'games',
'@brand:(sony|matias|beyerdynamic|(mad catz))', 'GROUPBY', 1,
'@brand', 'REDUCE', 'FIRST_VALUE', 4, '@title', 'BY', '@price',
'DESC', 'AS', 'top_item', 'REDUCE', 'FIRST_VALUE', 4, '@price',
'BY', '@price', 'DESC', 'AS', 'top_price', 'REDUCE', 'FIRST_VALUE',
4, '@title', 'BY', '@price', 'ASC', 'AS', 'bottom_item', 'REDUCE',
'FIRST_VALUE', 4, '@price', 'BY', '@price', 'ASC', 'AS',
'bottom_price', 'SORTBY', 2, '@top_price', 'DESC', 'MAX', 5)
expected = [
4L,
[
'brand', 'sony', 'top_item',
'sony psp slim & lite 2000 console', 'top_price', '695.8',
'bottom_item',
'sony dlchd20p high speed hdmi cable for playstation 3',
'bottom_price', '5.88'
],
[
'brand', 'matias', 'top_item', 'matias halfkeyboard usb',
'top_price', '559.99', 'bottom_item',
'matias halfkeyboard usb', 'bottom_price', '559.99'
],
[
'brand', 'beyerdynamic', 'top_item',
'beyerdynamic mmx300 pc gaming premium digital headset with microphone',
'top_price', '359.74', 'bottom_item',
'beyerdynamic headzone pc gaming digital surround sound system with mmx300 digital headset with microphone',
'bottom_price', '0'
],
[
'brand', 'mad catz', 'top_item',
'mad catz s.t.r.i.k.e.7 gaming keyboard', 'top_price',
'295.95', 'bottom_item',
'madcatz mov4545 xbox replacement breakaway cable',
'bottom_price', '3.49'
]
]
self.env.assertListEqual(expected, res)
def testLoadAfterGroupBy(self):
with self.env.assertResponseError():
self.env.cmd('ft.aggregate', 'games', '*', 'GROUPBY', 1, '@brand',
'LOAD', 1, '@brand')
class TestAggregate():
def __init__(self):
self.env = Env()
add_values(self.env)
def testGroupBy(self):
cmd = [
'ft.aggregate', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'count', '0', 'AS', 'count', 'SORTBY', 2, '@count', 'desc',
'LIMIT', '0', '5'
]
res = self.env.cmd(*cmd)
self.env.assertIsNotNone(res)
self.env.assertEqual([
292L, ['brand', '', 'count', '1518'],
['brand', 'mad catz', 'count', '43'],
['brand', 'generic', 'count', '40'],
['brand', 'steelseries', 'count', '37'],
['brand', 'logitech', 'count', '35']
], res)
def testMinMax(self):
cmd = [
'ft.aggregate', 'games', 'sony', 'GROUPBY', '1', '@brand',
'REDUCE', 'count', '0', 'REDUCE', 'min', '1', '@price', 'as',
'minPrice', 'SORTBY', '2', '@minPrice', 'DESC'
]
res = self.env.cmd(*cmd)
self.env.assertIsNotNone(res)
row = to_dict(res[1])
self.env.assertEqual(88, int(float(row['minPrice'])))
cmd = [
'ft.aggregate', 'games', 'sony', 'GROUPBY', '1', '@brand',
'REDUCE', 'count', '0', 'REDUCE', 'max', '1', '@price', 'as',
'maxPrice', 'SORTBY', '2', '@maxPrice', 'DESC'
]
res = self.env.cmd(*cmd)
row = to_dict(res[1])
self.env.assertEqual(695, int(float(row['maxPrice'])))
def testAvg(self):
cmd = [
'ft.aggregate', 'games', 'sony', 'GROUPBY', '1', '@brand',
'REDUCE', 'avg', '1', '@price', 'AS', 'avg_price', 'REDUCE',
'count', '0', 'SORTBY', '2', '@avg_price', 'DESC'
]
res = self.env.cmd(*cmd)
self.env.assertIsNotNone(res)
self.env.assertEqual(26, res[0])
# Ensure the formatting actually exists
first_row = to_dict(res[1])
self.env.assertEqual(109, int(float(first_row['avg_price'])))
for row in res[1:]:
row = to_dict(row)
self.env.assertIn('avg_price', row)
# Test aliasing
cmd = [
'FT.AGGREGATE', 'games', 'sony', 'GROUPBY', '1', '@brand',
'REDUCE', 'avg', '1', '@price', 'AS', 'avgPrice'
]
res = self.env.cmd(*cmd)
first_row = to_dict(res[1])
self.env.assertEqual(17, int(float(first_row['avgPrice'])))
def testCountDistinct(self):
cmd = [
'FT.AGGREGATE', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'COUNT_DISTINCT', '1', '@title', 'AS', 'count_distinct(title)',
'REDUCE', 'COUNT', '0'
]
res = self.env.cmd(*cmd)[1:]
# print res
row = to_dict(res[0])
self.env.assertEqual(1484, int(row['count_distinct(title)']))
cmd = [
'FT.AGGREGATE', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'COUNT_DISTINCTISH', '1', '@title', 'AS',
'count_distinctish(title)', 'REDUCE', 'COUNT', '0'
]
res = self.env.cmd(*cmd)[1:]
# print res
row = to_dict(res[0])
self.env.assertEqual(1461, int(row['count_distinctish(title)']))
def testQuantile(self):
cmd = [
'FT.AGGREGATE', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'QUANTILE', '2', '@price', '0.50', 'AS', 'q50', 'REDUCE',
'QUANTILE', '2', '@price', '0.90', 'AS', 'q90', 'REDUCE',
'QUANTILE', '2', '@price', '0.95', 'AS', 'q95', 'REDUCE', 'AVG',
'1', '@price', 'REDUCE', 'COUNT', '0', 'AS', 'rowcount', 'SORTBY',
'2', '@rowcount', 'DESC', 'MAX', '1'
]
res = self.env.cmd(*cmd)
row = to_dict(res[1])
# TODO: Better samples
self.env.assertAlmostEqual(14.99, float(row['q50']), delta=3)
self.env.assertAlmostEqual(70, float(row['q90']), delta=50)
self.env.assertAlmostEqual(110, (float(row['q95'])), delta=50)
def testStdDev(self):
cmd = [
'FT.AGGREGATE', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'STDDEV', '1', '@price', 'AS', 'stddev(price)', 'REDUCE', 'AVG',
'1', '@price', 'AS', 'avgPrice', 'REDUCE', 'QUANTILE', '2',
'@price', '0.50', 'AS', 'q50Price', 'REDUCE', 'COUNT', '0', 'AS',
'rowcount', 'SORTBY', '2', '@rowcount', 'DESC', 'LIMIT', '0', '10'
]
res = self.env.cmd(*cmd)
row = to_dict(res[1])
self.env.assertTrue(10 <= int(float(row['q50Price'])) <= 20)
self.env.assertAlmostEqual(53,
int(float(row['stddev(price)'])),
delta=50)
self.env.assertEqual(29, int(float(row['avgPrice'])))
def testParseTime(self):
cmd = [
'FT.AGGREGATE', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'COUNT', '0', 'AS', 'count', 'APPLY', 'timefmt(1517417144)', 'AS',
'dt', 'APPLY', 'parse_time("%FT%TZ", @dt)', 'as', 'parsed_dt',
'LIMIT', '0', '1'
]
res = self.env.cmd(*cmd)
self.env.assertEqual([
'brand', '', 'count', '1518', 'dt', '2018-01-31T16:45:44Z',
'parsed_dt', '1517417144'
], res[1])
def testRandomSample(self):
cmd = [
'FT.AGGREGATE', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'COUNT', '0', 'AS', 'num', 'REDUCE', 'RANDOM_SAMPLE', '2',
'@price', '10', 'SORTBY', '2', '@num', 'DESC', 'MAX', '10'
]
for row in self.env.cmd(*cmd)[1:]:
self.env.assertIsInstance(row[5], list)
self.env.assertGreater(len(row[5]), 0)
self.env.assertGreaterEqual(row[3], len(row[5]))
self.env.assertLessEqual(len(row[5]), 10)
def testTimeFunctions(self):
cmd = [
'FT.AGGREGATE', 'games', '*', 'APPLY', '1517417144', 'AS', 'dt',
'APPLY', 'timefmt(@dt)', 'AS', 'timefmt', 'APPLY', 'day(@dt)',
'AS', 'day', 'APPLY', 'hour(@dt)', 'AS', 'hour', 'APPLY',
'minute(@dt)', 'AS', 'minute', 'APPLY', 'month(@dt)', 'AS',
'month', 'APPLY', 'dayofweek(@dt)', 'AS', 'dayofweek', 'APPLY',
'dayofmonth(@dt)', 'AS', 'dayofmonth', 'APPLY', 'dayofyear(@dt)',
'AS', 'dayofyear', 'APPLY', 'year(@dt)', 'AS', 'year', 'LIMIT',
'0', '1'
]
res = self.env.cmd(*cmd)
self.env.assertListEqual([
1L,
[
'dt', '1517417144', 'timefmt', '2018-01-31T16:45:44Z', 'day',
'1517356800', 'hour', '1517414400', 'minute', '1517417100',
'month', '1514764800', 'dayofweek', '3', 'dayofmonth', '31',
'dayofyear', '30', 'year', '2018'
]
], res)
def testStringFormat(self):
cmd = [
'FT.AGGREGATE', 'games', '@brand:sony', 'GROUPBY', '2', '@title',
'@brand', 'REDUCE', 'COUNT', '0', 'REDUCE', 'MAX', '1', '@price',
'AS', 'price', 'APPLY',
'format("%s|%s|%s|%s", @title, @brand, "Mark", @price)', 'as',
'titleBrand', 'LIMIT', '0', '10'
]
res = self.env.cmd(*cmd)
for row in res[1:]:
row = to_dict(row)
expected = '%s|%s|%s|%g' % (row['title'], row['brand'], 'Mark',
float(row['price']))
self.env.assertEqual(expected, row['titleBrand'])
def testSum(self):
cmd = [
'ft.aggregate', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'count', '0', 'AS', 'count', 'REDUCE', 'sum', 1, '@price', 'AS',
'sum(price)', 'SORTBY', 2, '@sum(price)', 'desc', 'LIMIT', '0', '5'
]
res = self.env.cmd(*cmd)
self.env.assertEqual([
292L, ['brand', '', 'count', '1518', 'sum(price)', '44780.69'],
['brand', 'mad catz', 'count', '43', 'sum(price)', '3973.48'],
['brand', 'razer', 'count', '26', 'sum(price)', '2558.58'],
['brand', 'logitech', 'count', '35', 'sum(price)', '2329.21'],
['brand', 'steelseries', 'count', '37', 'sum(price)', '1851.12']
], res)
def testFilter(self):
cmd = [
'ft.aggregate', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'count', '0', 'AS', 'count', 'FILTER', '@count > 5'
]
res = self.env.cmd(*cmd)
for row in res[1:]:
row = to_dict(row)
self.env.assertGreater(int(row['count']), 5)
cmd = [
'ft.aggregate', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'count', '0', 'AS', 'count', 'FILTER', '@count < 5', 'FILTER',
'@count > 2 && @brand != ""'
]
res = self.env.cmd(*cmd)
for row in res[1:]:
row = to_dict(row)
self.env.assertLess(int(row['count']), 5)
self.env.assertGreater(int(row['count']), 2)
def testToList(self):
cmd = [
'ft.aggregate', 'games', '*', 'GROUPBY', '1', '@brand', 'REDUCE',
'count_distinct', '1', '@price', 'as', 'count', 'REDUCE', 'tolist',
1, '@price', 'as', 'prices', 'SORTBY', 2, '@count', 'desc',
'LIMIT', '0', '5'
]
res = self.env.cmd(*cmd)
for row in res[1:]:
row = to_dict(row)
self.env.assertEqual(int(row['count']), len(row['prices']))
def testSortBy(self):
res = self.env.cmd('ft.aggregate', 'games', '*', 'GROUPBY', '1',
'@brand', 'REDUCE', 'sum', 1, '@price', 'as',
'price', 'SORTBY', 2, '@price', 'desc', 'LIMIT',
'0', '2')
self.env.assertListEqual([
292L, ['brand', '', 'price', '44780.69'],
['brand', 'mad catz', 'price', '3973.48']
], res)
res = self.env.cmd('ft.aggregate', 'games', '*', 'GROUPBY', '1',
'@brand', 'REDUCE', 'sum', 1, '@price', 'as',
'price', 'SORTBY', 2, '@price', 'asc', 'LIMIT', '0',
'2')
self.env.assertListEqual([
292L, ['brand', 'myiico', 'price', '0.23'],
['brand', 'crystal dynamics', 'price', '0.25']
], res)
# Test MAX with limit higher than it
res = self.env.cmd('ft.aggregate', 'games', '*', 'GROUPBY', '1',
'@brand', 'REDUCE', 'sum', 1, '@price', 'as',
'price', 'SORTBY', 2, '@price', 'asc', 'MAX', 2)
self.env.assertListEqual([
292L, ['brand', 'myiico', 'price', '0.23'],
['brand', 'crystal dynamics', 'price', '0.25']
], res)
# Test Sorting by multiple properties
res = self.env.cmd(
'ft.aggregate',
'games',
'*',
'GROUPBY',
'1',
'@brand',
'REDUCE',
'sum',
1,
'@price',
'as',
'price',
'APPLY',
'(@price % 10)',
'AS',
'price',
'SORTBY',
4,
'@price',
'asc',
'@brand',
'desc',
'MAX',
10,
)
self.env.assertListEqual([
292L, ['brand', 'zps', 'price', '0'],
['brand', 'zalman', 'price', '0'], [
'brand', 'yoozoo', 'price', '0'
], ['brand', 'white label', 'price', '0'],
['brand', 'stinky', 'price', '0'],
['brand', 'polaroid', 'price', '0'],
['brand', 'plantronics', 'price', '0'],
['brand', 'ozone', 'price', '0'], ['brand', 'oooo', 'price', '0'],
['brand', 'neon', 'price', '0']
], res)
def testExpressions(self):
pass
def testNoGroup(self):
res = self.env.cmd(
'ft.aggregate',
'games',
'*',
'LOAD',
'2',
'@brand',
'@price',
'APPLY',
'floor(sqrt(@price)) % 10',
'AS',
'price',
'SORTBY',
4,
'@price',
'desc',
'@brand',
'desc',
'MAX',
5,
)
exp = [
2265L, ['brand', 'Xbox', 'price', '9'],
['brand', 'turtle beach', 'price', '9'],
['brand', 'trust', 'price', '9'],
['brand', 'steelseries', 'price', '9'],
['brand', 'speedlink', 'price', '9']
]
# exp = [2265L, ['brand', 'Xbox', 'price', '9'], ['brand', 'Turtle Beach', 'price', '9'], [
# 'brand', 'Trust', 'price', '9'], ['brand', 'SteelSeries', 'price', '9'], ['brand', 'Speedlink', 'price', '9']]
self.env.assertListEqual(exp[1], res[1])
def testLoad(self):
res = self.env.cmd('ft.aggregate', 'games', '*', 'LOAD', '3', '@brand',
'@price', '@nonexist', 'SORTBY', 2, '@price',
'DESC', 'MAX', 2)
exp = [
3L, ['brand', '', 'price', '759.12'],
['brand', 'Sony', 'price', '695.8']
]
self.env.assertEqual(exp[1], res[1])
self.env.assertEqual(exp[2], res[2])
def testLoadWithDocId(self):
res = self.env.cmd('ft.aggregate', 'games', '*', 'LOAD', '3', '@brand',
'@price', '@__key', 'SORTBY', 2, '@price', 'DESC',
'MAX', 4)
exp = [
3L, ['brand', '', 'price', '759.12', '__key', 'B00006JJIC'],
['brand', 'Sony', 'price', '695.8', '__key', 'B000F6W1AG']
]
self.env.assertEqual(exp[1], res[1])
self.env.assertEqual(exp[2], res[2])
res = self.env.cmd('ft.aggregate', 'games', '*', 'LOAD', '3', '@brand',
'@price', '@__key', 'FILTER',
'@__key == "B000F6W1AG"')
self.env.assertEqual(
res[1], ['brand', 'Sony', 'price', '695.8', '__key', 'B000F6W1AG'])
def testLoadImplicit(self):
# same as previous
res = self.env.cmd('ft.aggregate', 'games', '*', 'LOAD', '1', '@brand',
'SORTBY', 2, '@price', 'DESC')
exp = [
3L, ['brand', '', 'price', '759.12'],
['brand', 'Sony', 'price', '695.8']
]
self.env.assertEqual(exp[1], res[1])
def testSplit(self):
res = self.env.cmd(
'ft.aggregate', 'games', '*', 'APPLY',
'split("hello world, foo,,,bar,", ",", " ")', 'AS', 'strs',
'APPLY', 'split("hello world, foo,,,bar,", " ", ",")', 'AS',
'strs2', 'APPLY', 'split("hello world, foo,,,bar,", "", "")',
'AS', 'strs3', 'APPLY', 'split("hello world, foo,,,bar,")', 'AS',
'strs4', 'APPLY', 'split("hello world, foo,,,bar,",",")', 'AS',
'strs5', 'APPLY', 'split("")', 'AS', 'empty', 'LIMIT', '0', '1')
# print "Got {} results".format(len(res))
# return
# pprint.pprint(res)
self.env.assertListEqual([
1L,
[
'strs', ['hello world', 'foo', 'bar'], 'strs2',
['hello', 'world', 'foo,,,bar'], 'strs3',
['hello world, foo,,,bar,'], 'strs4',
['hello world', 'foo', 'bar'], 'strs5',
['hello world', 'foo', 'bar'], 'empty', []
]
], res)
def testFirstValue(self):
res = self.env.cmd(
'ft.aggregate', 'games',
'@brand:(sony|matias|beyerdynamic|(mad catz))', 'GROUPBY', 1,
'@brand', 'REDUCE', 'FIRST_VALUE', 4, '@title', 'BY', '@price',
'DESC', 'AS', 'top_item', 'REDUCE', 'FIRST_VALUE', 4, '@price',
'BY', '@price', 'DESC', 'AS', 'top_price', 'REDUCE', 'FIRST_VALUE',
4, '@title', 'BY', '@price', 'ASC', 'AS', 'bottom_item', 'REDUCE',
'FIRST_VALUE', 4, '@price', 'BY', '@price', 'ASC', 'AS',
'bottom_price', 'SORTBY', 2, '@top_price', 'DESC', 'MAX', 5)
expected = [
4L,
[
'brand', 'sony', 'top_item',
'sony psp slim & lite 2000 console', 'top_price', '695.8',
'bottom_item',
'sony dlchd20p high speed hdmi cable for playstation 3',
'bottom_price', '5.88'
],
[
'brand', 'matias', 'top_item', 'matias halfkeyboard usb',
'top_price', '559.99', 'bottom_item',
'matias halfkeyboard usb', 'bottom_price', '559.99'
],
[
'brand', 'beyerdynamic', 'top_item',
'beyerdynamic mmx300 pc gaming premium digital headset with microphone',
'top_price', '359.74', 'bottom_item',
'beyerdynamic headzone pc gaming digital surround sound system with mmx300 digital headset with microphone',
'bottom_price', '0'
],
[
'brand', 'mad catz', 'top_item',
'mad catz s.t.r.i.k.e.7 gaming keyboard', 'top_price',
'295.95', 'bottom_item',
'madcatz mov4545 xbox replacement breakaway cable',
'bottom_price', '3.49'
]
]
# hack :(
def mklower(result):
for arr in result[1:]:
for x in range(len(arr)):
arr[x] = arr[x].lower()
mklower(expected)
mklower(res)
self.env.assertListEqual(expected, res)
def testLoadAfterGroupBy(self):
with self.env.assertResponseError():
self.env.cmd('ft.aggregate', 'games', '*', 'GROUPBY', 1, '@brand',
'LOAD', 1, '@brand')
def testReducerGeneratedAliasing(self):
rv = self.env.cmd('ft.aggregate', 'games', '*', 'GROUPBY', 1, '@brand',
'REDUCE', 'MIN', 1, '@price', 'LIMIT', 0, 1)
self.env.assertEqual(
[292L, ['brand', '', '__generated_aliasminprice', '0']], rv)
rv = self.env.cmd('ft.aggregate', 'games',
'@brand:(sony|matias|beyerdynamic|(mad catz))',
'GROUPBY', 1, '@brand', 'REDUCE', 'FIRST_VALUE', 4,
'@title', 'BY', '@price', 'DESC', 'SORTBY', 2,
'@brand', 'ASC')
self.env.assertEqual('__generated_aliasfirst_valuetitle,by,price,desc',
rv[1][2])
def testIssue1125(self):
self.env.skipOnCluster()
# SEARCH should fail
self.env.expect('ft.search', 'games', '*', 'limit', 0, 2000000).error() \
.contains('LIMIT exceeds maximum of 1000000')
# SEARCH should succeed
self.env.expect('ft.config', 'set', 'MAXSEARCHRESULTS', -1).ok()
rv = self.env.cmd('ft.search', 'games', '*', 'LIMIT', 0, 12345678)
self.env.assertEqual(4531, len(rv))
# AGGREGATE should succeed
rv = self.env.cmd('ft.aggregate', 'games', '*', 'LIMIT', 0, 12345678)
self.env.assertEqual(2266, len(rv))
# AGGREGATE should fail
self.env.expect('ft.config', 'set', 'MAXAGGREGATERESULTS',
1000000).ok()
self.env.expect('ft.aggregate', 'games', '*', 'limit', 0, 2000000).error() \
.contains('LIMIT exceeds maximum of 1000000')
# force global limit on aggregate
num = 10
self.env.expect('ft.config', 'set', 'MAXAGGREGATERESULTS', num).ok()
rv = self.env.cmd('ft.aggregate', 'games', '*')
self.env.assertEqual(num + 1, len(rv))
def testMultiSortBy(self):
self.env.expect('ft.aggregate', 'games', '*',
'LOAD', '2', '@brand', '@price',
'SORTBY', 2, '@brand', 'DESC',
'SORTBY', 2, '@price', 'DESC').error()\
.contains('Multiple SORTBY steps are not allowed. Sort multiple fields in a single step')
class testIndexUpdatesFlow(FlowTestsBase):
def __init__(self):
self.env = Env()
global redis_graph
redis_con = self.env.getConnection()
redis_graph = Graph(GRAPH_ID, redis_con)
self.populate_graph()
self.build_indices()
def new_node(self):
return Node(label=labels[node_ctr % 2],
properties={
'unique':
node_ctr,
'group':
random.choice(groups),
'doubleval':
round(random.uniform(-1, 1), 2),
'intval':
random.randint(1, 10000),
'stringval':
''.join(
random.choice(string.lowercase) for x in range(6))
})
def populate_graph(self):
global node_ctr
for i in range(1000):
node = self.new_node()
redis_graph.add_node(node)
node_ctr += 1
redis_graph.commit()
def build_indices(self):
for field in fields:
redis_graph.redis_con.execute_command(
"GRAPH.QUERY", GRAPH_ID,
"CREATE INDEX ON :label_a(%s)" % (field))
redis_graph.redis_con.execute_command(
"GRAPH.QUERY", GRAPH_ID,
"CREATE INDEX ON :label_b(%s)" % (field))
# Validate that all properties are indexed
def validate_indexed(self):
for field in fields:
resp = redis_graph.execution_plan(
"""MATCH (a:label_a) WHERE a.%s > 0 RETURN a""" % (field))
self.env.assertIn('Index Scan', resp)
# So long as 'unique' is not modified, label_a.unique will always be even and label_b.unique will always be odd
def validate_unique(self):
result = redis_graph.query("MATCH (a:label_a) RETURN a.unique")
# Remove the header
result.result_set.pop(0)
for val in result.result_set:
self.env.assertEquals(int(float(val[0])) % 2, 0)
result = redis_graph.query("MATCH (b:label_b) RETURN b.unique")
# Remove the header
result.result_set.pop(0)
for val in result.result_set:
self.env.assertEquals(int(float(val[0])) % 2, 1)
# The index scan ought to return identical results to a label scan over the same range of values.
def validate_doubleval(self):
for label in labels:
resp = redis_graph.execution_plan(
"""MATCH (a:%s) WHERE a.doubleval < 100 RETURN a.doubleval ORDER BY a.doubleval"""
% (label))
self.env.assertIn('Index Scan', resp)
indexed_result = redis_graph.query(
"""MATCH (a:%s) WHERE a.doubleval < 100 RETURN a.doubleval ORDER BY a.doubleval"""
% (label))
scan_result = redis_graph.query(
"""MATCH (a:%s) RETURN a.doubleval ORDER BY a.doubleval""" %
(label))
self.env.assertEqual(len(indexed_result.result_set),
len(scan_result.result_set))
# Collect any elements between the two result sets that fail a string comparison
# so that we may compare them as doubles (specifically, -0 and 0 should be considered equal)
differences = [[i[0], j[0]] for i, j in zip(
indexed_result.result_set, scan_result.result_set) if i != j]
for pair in differences:
self.env.assertEqual(float(pair[0]), float(pair[1]))
# The intval property can be assessed similar to doubleval, but the result sets should be identical
def validate_intval(self):
for label in labels:
resp = redis_graph.execution_plan(
"""MATCH (a:%s) WHERE a.intval > 0 RETURN a.intval ORDER BY a.intval"""
% (label))
self.env.assertIn('Index Scan', resp)
indexed_result = redis_graph.query(
"""MATCH (a:%s) WHERE a.intval > 0 RETURN a.intval ORDER BY a.intval"""
% (label))
scan_result = redis_graph.query(
"""MATCH (a:%s) RETURN a.intval ORDER BY a.intval""" % (label))
self.env.assertEqual(indexed_result.result_set,
scan_result.result_set)
# Validate a series of premises to ensure that the graph has not been modified unexpectedly
def validate_state(self):
self.validate_unique()
self.validate_indexed()
self.validate_doubleval()
self.validate_intval()
# Modify a property, triggering updates to all nodes in two indices
def test01_full_property_update(self):
result = redis_graph.query(
"MATCH (a) SET a.doubleval = a.doubleval + %f" %
(round(random.uniform(-1, 1), 2)))
self.env.assertEquals(result.properties_set, 1000)
# Verify that index scans still function and return correctly
self.validate_state()
# Modify a property, triggering updates to a subset of nodes in two indices
def test02_partial_property_update(self):
redis_graph.query(
"MATCH (a) WHERE a.doubleval > 0 SET a.doubleval = a.doubleval + %f"
% (round(random.uniform(-1, 1), 2)))
# Verify that index scans still function and return correctly
self.validate_state()
# Add 100 randomized nodes and validate indices
def test03_node_creation(self):
# Reset nodes in the Graph object so that we won't double-commit the originals
redis_graph.nodes = {}
global node_ctr
for i in range(100):
node = self.new_node()
redis_graph.add_node(node)
node_ctr += 1
redis_graph.commit()
self.validate_state()
# Delete every other node in first 100 and validate indices
def test04_node_deletion(self):
# Reset nodes in the Graph object so that we won't double-commit the originals
redis_graph.nodes = {}
global node_ctr
# Delete nodes one at a time
for i in range(0, 100, 2):
result = redis_graph.query("MATCH (a) WHERE ID(a) = %d DELETE a" %
(i))
self.env.assertEquals(result.nodes_deleted, 1)
node_ctr -= 1
self.validate_state()
# Delete all nodes matching a filter
result = redis_graph.query(
"MATCH (a:label_a) WHERE a.group = 'Group A' DELETE a")
self.env.assertGreater(result.nodes_deleted, 0)
self.validate_state()
class testGraphBulkInsertFlow(FlowTestsBase):
def __init__(self):
self.env = Env(decodeResponses=True)
global redis_graph
global redis_con
redis_con = self.env.getConnection()
redis_graph = Graph("graph", redis_con)
# Run bulk loader script and validate terminal output
def test01_run_script(self):
graphname = "graph"
runner = CliRunner()
csv_path = os.path.dirname(os.path.abspath(
__file__)) + '/../../demo/social/resources/bulk_formatted/'
res = runner.invoke(bulk_insert, [
'--port', port, '--nodes', csv_path + 'Person.csv', '--nodes',
csv_path + 'Country.csv', '--relations', csv_path + 'KNOWS.csv',
'--relations', csv_path + 'VISITED.csv', graphname
])
# The script should report 27 node creations and 48 edge creations
self.env.assertEquals(res.exit_code, 0)
self.env.assertIn('27 nodes created', res.output)
self.env.assertIn('56 relations created', res.output)
# Validate that the expected nodes and properties have been constructed
def test02_validate_nodes(self):
global redis_graph
# Query the newly-created graph
query_result = redis_graph.query(
'MATCH (p:Person) RETURN p.name, p.age, p.gender, p.status, ID(p) ORDER BY p.name'
)
# Verify that the Person label exists, has the correct attributes, and is properly populated
expected_result = [['Ailon Velger', 32, 'male', 'married', 2],
['Alon Fital', 32, 'male', 'married', 1],
['Boaz Arad', 31, 'male', 'married', 4],
['Gal Derriere', 26, 'male', 'single', 11],
['Jane Chernomorin', 31, 'female', 'married', 8],
['Lucy Yanfital', 30, 'female', 'married', 7],
['Mor Yesharim', 31, 'female', 'married', 12],
['Noam Nativ', 34, 'male', 'single', 13],
['Omri Traub', 33, 'male', 'single', 5],
['Ori Laslo', 32, 'male', 'married', 3],
['Roi Lipman', 32, 'male', 'married', 0],
['Shelly Laslo Rooz', 31, 'female', 'married', 9],
['Tal Doron', 32, 'male', 'single', 6],
[
'Valerie Abigail Arad', 31, 'female', 'married',
10
]]
self.env.assertEquals(query_result.result_set, expected_result)
# Verify that the Country label exists, has the correct attributes, and is properly populated
query_result = redis_graph.query(
'MATCH (c:Country) RETURN c.name, ID(c) ORDER BY c.name')
expected_result = [['Andora', 21], ['Canada', 18], ['China', 19],
['Germany', 24], ['Greece', 17], ['Italy', 25],
['Japan', 16], ['Kazakhstan', 22],
['Netherlands', 20], ['Prague', 15], ['Russia', 23],
['Thailand', 26], ['USA', 14]]
self.env.assertEquals(query_result.result_set, expected_result)
# Validate that the expected relations and properties have been constructed
def test03_validate_relations(self):
# Query the newly-created graph
query_result = redis_graph.query(
'MATCH (a)-[e:KNOWS]->(b) RETURN a.name, e.relation, b.name ORDER BY e.relation, a.name, b.name'
)
expected_result = [['Ailon Velger', 'friend', 'Noam Nativ'],
['Alon Fital', 'friend', 'Gal Derriere'],
['Alon Fital', 'friend', 'Mor Yesharim'],
['Boaz Arad', 'friend', 'Valerie Abigail Arad'],
['Roi Lipman', 'friend', 'Ailon Velger'],
['Roi Lipman', 'friend', 'Alon Fital'],
['Roi Lipman', 'friend', 'Boaz Arad'],
['Roi Lipman', 'friend', 'Omri Traub'],
['Roi Lipman', 'friend', 'Ori Laslo'],
['Roi Lipman', 'friend', 'Tal Doron'],
['Ailon Velger', 'married', 'Jane Chernomorin'],
['Alon Fital', 'married', 'Lucy Yanfital'],
['Ori Laslo', 'married', 'Shelly Laslo Rooz']]
self.env.assertEquals(query_result.result_set, expected_result)
query_result = redis_graph.query(
'MATCH (a)-[e:VISITED]->(b) RETURN a.name, e.purpose, b.name ORDER BY e.purpose, a.name, b.name'
)
expected_result = [['Alon Fital', 'business', 'Prague'],
['Alon Fital', 'business', 'USA'],
['Boaz Arad', 'business', 'Netherlands'],
['Boaz Arad', 'business', 'USA'],
['Gal Derriere', 'business', 'Netherlands'],
['Jane Chernomorin', 'business', 'USA'],
['Lucy Yanfital', 'business', 'USA'],
['Mor Yesharim', 'business', 'Germany'],
['Ori Laslo', 'business', 'China'],
['Ori Laslo', 'business', 'USA'],
['Roi Lipman', 'business', 'Prague'],
['Roi Lipman', 'business', 'USA'],
['Tal Doron', 'business', 'Japan'],
['Tal Doron', 'business', 'USA'],
['Alon Fital', 'pleasure', 'Greece'],
['Alon Fital', 'pleasure', 'Prague'],
['Alon Fital', 'pleasure', 'USA'],
['Boaz Arad', 'pleasure', 'Netherlands'],
['Boaz Arad', 'pleasure', 'USA'],
['Jane Chernomorin', 'pleasure', 'Greece'],
['Jane Chernomorin', 'pleasure', 'Netherlands'],
['Jane Chernomorin', 'pleasure', 'USA'],
['Lucy Yanfital', 'pleasure', 'Kazakhstan'],
['Lucy Yanfital', 'pleasure', 'Prague'],
['Lucy Yanfital', 'pleasure', 'USA'],
['Mor Yesharim', 'pleasure', 'Greece'],
['Mor Yesharim', 'pleasure', 'Italy'],
['Noam Nativ', 'pleasure', 'Germany'],
['Noam Nativ', 'pleasure', 'Netherlands'],
['Noam Nativ', 'pleasure', 'Thailand'],
['Omri Traub', 'pleasure', 'Andora'],
['Omri Traub', 'pleasure', 'Greece'],
['Omri Traub', 'pleasure', 'USA'],
['Ori Laslo', 'pleasure', 'Canada'],
['Roi Lipman', 'pleasure', 'Japan'],
['Roi Lipman', 'pleasure', 'Prague'],
['Shelly Laslo Rooz', 'pleasure', 'Canada'],
['Shelly Laslo Rooz', 'pleasure', 'China'],
['Shelly Laslo Rooz', 'pleasure', 'USA'],
['Tal Doron', 'pleasure', 'Andora'],
['Tal Doron', 'pleasure', 'USA'],
['Valerie Abigail Arad', 'pleasure', 'Netherlands'],
['Valerie Abigail Arad', 'pleasure', 'Russia']]
self.env.assertEquals(query_result.result_set, expected_result)
def test04_private_identifiers(self):
graphname = "tmpgraph1"
# Write temporary files
with open('/tmp/nodes.tmp', mode='w') as csv_file:
out = csv.writer(csv_file)
out.writerow(["_identifier", "nodename"])
out.writerow([0, "a"])
out.writerow([5, "b"])
out.writerow([3, "c"])
with open('/tmp/relations.tmp', mode='w') as csv_file:
out = csv.writer(csv_file)
out.writerow(["src", "dest"])
out.writerow([0, 3])
out.writerow([5, 3])
runner = CliRunner()
res = runner.invoke(bulk_insert, [
'--port', port, '--nodes', '/tmp/nodes.tmp', '--relations',
'/tmp/relations.tmp', graphname
])
# The script should report 3 node creations and 2 edge creations
self.env.assertEquals(res.exit_code, 0)
self.env.assertIn('3 nodes created', res.output)
self.env.assertIn('2 relations created', res.output)
# Delete temporary files
os.remove('/tmp/nodes.tmp')
os.remove('/tmp/relations.tmp')
tmp_graph = Graph(graphname, redis_con)
# The field "_identifier" should not be a property in the graph
query_result = tmp_graph.query('MATCH (a) RETURN a')
for propname in query_result.header:
self.env.assertNotIn('_identifier', propname)
def test05_reused_identifier(self):
graphname = "tmpgraph2"
# Write temporary files
with open('/tmp/nodes.tmp', mode='w') as csv_file:
out = csv.writer(csv_file)
out.writerow(["_identifier", "nodename"])
out.writerow([0, "a"])
out.writerow([5, "b"])
out.writerow([0, "c"]) # reused identifier
with open('/tmp/relations.tmp', mode='w') as csv_file:
out = csv.writer(csv_file)
out.writerow(["src", "dest"])
out.writerow([0, 3])
runner = CliRunner()
res = runner.invoke(bulk_insert, [
'--port', port, '--nodes', '/tmp/nodes.tmp', '--relations',
'/tmp/relations.tmp', graphname
])
# The script should fail because a node identifier is reused
self.env.assertNotEqual(res.exit_code, 0)
self.env.assertIn('used multiple times', res.output)
# Run the script again without creating relations
runner = CliRunner()
res = runner.invoke(
bulk_insert,
['--port', port, '--nodes', '/tmp/nodes.tmp', graphname])
# The script should succeed and create 3 nodes
self.env.assertEquals(res.exit_code, 0)
self.env.assertIn('3 nodes created', res.output)
# Delete temporary files
os.remove('/tmp/nodes.tmp')
os.remove('/tmp/relations.tmp')
def test06_batched_build(self):
# Create demo graph wth one query per input file
graphname = "batched_graph"
runner = CliRunner()
csv_path = os.path.dirname(os.path.abspath(
__file__)) + '/../../demo/social/resources/bulk_formatted/'
res = runner.invoke(bulk_insert, [
'--port', port, '--nodes', csv_path + 'Person.csv', '--nodes',
csv_path + 'Country.csv', '--relations', csv_path + 'KNOWS.csv',
'--relations', csv_path + 'VISITED.csv', '--max-token-count', 1,
graphname
])
self.env.assertEquals(res.exit_code, 0)
# The script should report statistics multiple times
self.env.assertGreater(res.output.count('nodes created'), 1)
new_graph = Graph(graphname, redis_con)
# Newly-created graph should be identical to graph created in single query
original_result = redis_graph.query(
'MATCH (p:Person) RETURN p, ID(p) ORDER BY p.name')
new_result = new_graph.query(
'MATCH (p:Person) RETURN p, ID(p) ORDER BY p.name')
self.env.assertEquals(original_result.result_set,
new_result.result_set)
original_result = redis_graph.query(
'MATCH (a)-[e:KNOWS]->(b) RETURN a.name, e, b.name ORDER BY e.relation, a.name'
)
new_result = new_graph.query(
'MATCH (a)-[e:KNOWS]->(b) RETURN a.name, e, b.name ORDER BY e.relation, a.name'
)
self.env.assertEquals(original_result.result_set,
new_result.result_set)
def test07_script_failures(self):
graphname = "tmpgraph3"
# Write temporary files
with open('/tmp/nodes.tmp', mode='w') as csv_file:
out = csv.writer(csv_file)
out.writerow(["id", "nodename"])
out.writerow([0]) # Wrong number of properites
runner = CliRunner()
res = runner.invoke(
bulk_insert,
['--port', port, '--nodes', '/tmp/nodes.tmp', graphname])
# The script should fail because a row has the wrong number of fields
self.env.assertNotEqual(res.exit_code, 0)
self.env.assertIn('Expected 2 columns', str(res.exception))
# Write temporary files
with open('/tmp/nodes.tmp', mode='w') as csv_file:
out = csv.writer(csv_file)
out.writerow(["id", "nodename"])
out.writerow([0, "a"])
with open('/tmp/relations.tmp', mode='w') as csv_file:
out = csv.writer(csv_file)
out.writerow(["src"]) # Incomplete relation description
out.writerow([0])
runner = CliRunner()
res = runner.invoke(bulk_insert, [
'--port', port, '--nodes', '/tmp/nodes.tmp', '--relations',
'/tmp/relations.tmp', graphname
])
# The script should fail because a row has the wrong number of fields
self.env.assertNotEqual(res.exit_code, 0)
self.env.assertIn('should have at least 2 elements',
str(res.exception))
with open('/tmp/relations.tmp', mode='w') as csv_file:
out = csv.writer(csv_file)
out.writerow(["src", "dest"])
out.writerow([0, "fakeidentifier"])
runner = CliRunner()
res = runner.invoke(bulk_insert, [
'--port', port, '--nodes', '/tmp/nodes.tmp', '--relations',
'/tmp/relations.tmp', graphname
])
# The script should fail because an invalid node identifier was used
self.env.assertNotEqual(res.exit_code, 0)
self.env.assertIn('fakeidentifier', str(res.exception))
os.remove('/tmp/nodes.tmp')
os.remove('/tmp/relations.tmp')
# Test passing invalid arguments directly to the GRAPH.BULK endpoint
try:
redis_con.execute_command("GRAPH.BULK", "a", "a", "a")
self.env.assertTrue(False)
except redis.exceptions.ResponseError as e:
self.env.assertIn("Invalid graph operation on empty key", str(e))
# Verify that numeric, boolean, and null types are properly handled
def test08_property_types(self):
graphname = "tmpgraph4"
# Write temporary files
with open('/tmp/nodes.tmp', mode='w') as csv_file:
out = csv.writer(csv_file)
out.writerow(["numeric", "mixed", "bool"])
out.writerow([0, '', True])
out.writerow([5, "notnull", False])
out.writerow([7, '', False]) # reused identifier
with open('/tmp/relations.tmp', mode='w') as csv_file:
out = csv.writer(csv_file)
out.writerow(["src", "dest", "prop"])
out.writerow([0, 5, True])
out.writerow([5, 7, 3.5])
out.writerow([7, 0, ''])
runner = CliRunner()
res = runner.invoke(bulk_insert, [
'--port', port, '--nodes', '/tmp/nodes.tmp', '--relations',
'/tmp/relations.tmp', graphname
])
self.env.assertEquals(res.exit_code, 0)
self.env.assertIn('3 nodes created', res.output)
self.env.assertIn('3 relations created', res.output)
graph = Graph(graphname, redis_con)
query_result = graph.query(
'MATCH (a)-[e]->() RETURN a.numeric, a.mixed, a.bool, e.prop ORDER BY a.numeric, e.prop'
)
expected_result = [[0, None, True, True], [5, 'notnull', False, 3.5],
[7, None, False, None]]
# The graph should have the correct types for all properties
self.env.assertEquals(query_result.result_set, expected_result)
# Verify that the bulk loader does not block the server
def test09_large_bulk_insert(self):
graphname = "tmpgraph5"
prop_str = "Property value to be repeated 1 million generating a multi-megabyte CSV"
# Write temporary files
filename = '/tmp/nodes.tmp'
with open(filename, mode='w') as csv_file:
out = csv.writer(csv_file)
out.writerow(["long_property_string"])
for i in range(100_000):
out.writerow([prop_str])
# Instantiate a thread to run the bulk loader
thread = threading.Thread(target=run_bulk_loader,
args=(graphname, filename))
thread.start()
# Ping server while bulk-loader is running
ping_count = 0
while thread.is_alive():
t0 = time.time()
redis_con.ping()
t1 = time.time() - t0
# Verify that pinging the server takes less than 1 second during bulk insertion
self.env.assertLess(t1, 2)
ping_count += 1
thread.join()
# Verify that at least one ping was issued
self.env.assertGreater(ping_count, 1)
class testGraphDeletionFlow(FlowTestsBase):
def __init__(self):
self.env = Env()
global redis_graph
redis_con = self.env.getConnection()
redis_graph = Graph(GRAPH_ID, redis_con)
self.populate_graph()
def populate_graph(self):
nodes = {}
# Create entities
people = ["Roi", "Alon", "Ailon", "Boaz", "Tal", "Omri", "Ori"]
for p in people:
node = Node(label="person", properties={"name": p})
redis_graph.add_node(node)
nodes[p] = node
# Fully connected graph
for src in nodes:
for dest in nodes:
if src != dest:
edge = Edge(nodes[src], "know", nodes[dest])
redis_graph.add_edge(edge)
# Connect Roi to Alon via another edge type.
edge = Edge(nodes["Roi"], "SameBirthday", nodes["Alon"])
redis_graph.add_edge(edge)
redis_graph.commit()
# Count how many nodes contains the `name` attribute
# remove the `name` attribute from some nodes
# make sure the count updates accordingly,
# restore `name` attribute from, verify that count returns to its original value.
def test01_delete_attribute(self):
# How many nodes contains the 'name' attribute
query = """MATCH (n) WHERE EXISTS(n.name)=true RETURN count(n)"""
actual_result = redis_graph.query(query)
nodeCount = actual_result.result_set[0][0]
self.env.assertEquals(nodeCount, 7)
# Remove Tal's name attribute.
query = """MATCH (n) WHERE n.name = 'Tal' SET n.name = NULL"""
redis_graph.query(query)
# How many nodes contains the 'name' attribute,
# should reduce by 1 from previous count.
query = """MATCH (n) WHERE EXISTS(n.name)=true RETURN count(n)"""
actual_result = redis_graph.query(query)
nodeCount = actual_result.result_set[0][0]
self.env.assertEquals(nodeCount, 6)
# Reintroduce Tal's name attribute.
query = """MATCH (n) WHERE EXISTS(n.name)=false SET n.name = 'Tal'"""
actual_result = redis_graph.query(query)
# How many nodes contains the 'name' attribute
query = """MATCH (n) WHERE EXISTS(n.name)=true RETURN count(n)"""
actual_result = redis_graph.query(query)
nodeCount = actual_result.result_set[0][0]
self.env.assertEquals(nodeCount, 7)
# Delete edges pointing into either Boaz or Ori.
def test02_delete_edges(self):
query = """MATCH (s:person)-[e:know]->(d:person) WHERE d.name = "Boaz" OR d.name = "Ori" RETURN count(e)"""
actual_result = redis_graph.query(query)
edge_count = actual_result.result_set[0][0]
query = """MATCH (s:person)-[e:know]->(d:person) WHERE d.name = "Boaz" OR d.name = "Ori" DELETE e"""
actual_result = redis_graph.query(query)
self.env.assertEquals(actual_result.relationships_deleted, edge_count)
self.env.assertEquals(actual_result.nodes_deleted, 0)
# Make sure there are no edges going into either Boaz or Ori.
def test03_verify_edge_deletion(self):
query = """MATCH (s:person)-[e:know]->(d:person)
WHERE d.name = "Boaz" AND d.name = "Ori"
RETURN COUNT(s)"""
actual_result = redis_graph.query(query)
self.env.assertEquals(len(actual_result.result_set), 0)
# Remove 'know' edge connecting Roi to Alon
# Leaving a single edge of type SameBirthday
# connecting the two.
def test04_delete_typed_edge(self):
query = """MATCH (s:person {name: "Roi"})-[e:know]->(d:person {name: "Alon"})
RETURN count(e)"""
actual_result = redis_graph.query(query)
edge_count = actual_result.result_set[0][0]
query = """MATCH (s:person {name: "Roi"})-[e:know]->(d:person {name: "Alon"})
DELETE e"""
actual_result = redis_graph.query(query)
self.env.assertEquals(actual_result.relationships_deleted, edge_count)
self.env.assertEquals(actual_result.nodes_deleted, 0)
# Make sure Roi is still connected to Alon
# via the "SameBirthday" type edge.
def test05_verify_delete_typed_edge(self):
query = """MATCH (s:person {name: "Roi"})-[e:SameBirthday]->(d:person {name: "Alon"})
RETURN COUNT(s)"""
actual_result = redis_graph.query(query)
self.env.assertEquals(len(actual_result.result_set), 1)
query = """MATCH (s:person {name: "Roi"})-[e:know]->(d:person {name: "Alon"})
RETURN COUNT(s)"""
actual_result = redis_graph.query(query)
self.env.assertEquals(len(actual_result.result_set), 0)
# Remove both Alon and Boaz from the graph.
def test06_delete_nodes(self):
rel_count_query = """MATCH (a:person)-[e]->(b:person)
WHERE a.name = 'Boaz' OR a.name = 'Alon'
OR b.name = 'Boaz' OR b.name = 'Alon'
RETURN COUNT(e)"""
rel_count_result = redis_graph.query(rel_count_query)
# Get the total number of unique edges (incoming and outgoing)
# connected to Alon and Boaz.
rel_count = rel_count_result.result_set[0][0]
query = """MATCH (s:person)
WHERE s.name = "Boaz" OR s.name = "Alon"
DELETE s"""
actual_result = redis_graph.query(query)
self.env.assertEquals(actual_result.relationships_deleted, rel_count)
self.env.assertEquals(actual_result.nodes_deleted, 2)
# Make sure Alon and Boaz are not in the graph.
def test07_get_deleted_nodes(self):
query = """MATCH (s:person)
WHERE s.name = "Boaz" OR s.name = "Alon"
RETURN s"""
actual_result = redis_graph.query(query)
self.env.assertEquals(len(actual_result.result_set), 0)
# Make sure Alon and Boaz are the only removed nodes.
def test08_verify_node_deletion(self):
query = """MATCH (s:person)
RETURN COUNT(s)"""
actual_result = redis_graph.query(query)
nodeCount = actual_result.result_set[0][0]
self.env.assertEquals(nodeCount, 5)
def test09_delete_entire_graph(self):
# Make sure graph exists.
query = """MATCH (n) RETURN COUNT(n)"""
result = redis_graph.query(query)
nodeCount = result.result_set[0][0]
self.env.assertGreater(nodeCount, 0)
# Delete graph.
redis_graph.delete()
# Try to query a deleted graph.
redis_graph.query(query)
result = redis_graph.query(query)
nodeCount = result.result_set[0][0]
self.env.assertEquals(nodeCount, 0)
def test10_bulk_edge_deletion_timing(self):
# Create large amount of relationships (50000).
redis_graph.query(
"""UNWIND(range(1, 50000)) as x CREATE ()-[:R]->()""")
# Delete and benchmark for 300ms.
query = """MATCH (a)-[e:R]->(b) DELETE e"""
result = redis_graph.query(query)
query_info = QueryInfo(
query=query,
description=
"Test the execution time for deleting large number of edges",
max_run_time_ms=300)
# Test will not fail for execution time > 300ms but a warning will be shown at the test output.
self.env.assertEquals(result.relationships_deleted, 50000)
self._assert_run_time(result, query_info)
def test11_delete_entity_type_validation(self):
# Currently we only support deletion of either nodes or edges
# we've yet to introduce deletion of Path.
# Try to delete an integer.
query = """UNWIND [1] AS x DELETE x"""
try:
redis_graph.query(query)
self.env.assertTrue(False)
except Exception as error:
self.env.assertTrue("Delete type mismatch" in error.message)
query = """MATCH p=(n) DELETE p"""
try:
redis_graph.query(query)
self.env.assertTrue(False)
except Exception as error:
self.env.assertTrue("Delete type mismatch" in error.message)
def test12_delete_unwind_entity(self):
redis_con = self.env.getConnection()
redis_graph = Graph("delete_test", redis_con)
# Create 10 nodes.
for i in range(10):
redis_graph.add_node(Node())
redis_graph.flush()
# Unwind path nodes.
query = """MATCH p = () UNWIND nodes(p) AS node DELETE node"""
actual_result = redis_graph.query(query)
self.env.assertEquals(actual_result.nodes_deleted, 10)
self.env.assertEquals(actual_result.relationships_deleted, 0)
for i in range(10):
redis_graph.add_node(Node())
redis_graph.flush()
# Unwind collected nodes.
query = """MATCH (n) WITH collect(n) AS nodes UNWIND nodes AS node DELETE node"""
actual_result = redis_graph.query(query)
self.env.assertEquals(actual_result.nodes_deleted, 10)
self.env.assertEquals(actual_result.relationships_deleted, 0)
def test13_delete_path_elements(self):
self.env.flush()
redis_con = self.env.getConnection()
redis_graph = Graph("delete_test", redis_con)
src = Node()
dest = Node()
edge = Edge(src, "R", dest)
redis_graph.add_node(src)
redis_graph.add_node(dest)
redis_graph.add_edge(edge)
redis_graph.flush()
# Delete projected
# Unwind path nodes.
query = """MATCH p = (src)-[e]->(dest) WITH nodes(p)[0] AS node, relationships(p)[0] as edge DELETE node, edge"""
actual_result = redis_graph.query(query)
self.env.assertEquals(actual_result.nodes_deleted, 1)
self.env.assertEquals(actual_result.relationships_deleted, 1)
# Verify that variable-length traversals in each direction produce the correct results after deletion.
def test14_post_deletion_traversal_directions(self):
self.env.flush()
redis_con = self.env.getConnection()
redis_graph = Graph("G", redis_con)
nodes = {}
# Create entities.
labels = ["Dest", "Src", "Src2"]
for idx, l in enumerate(labels):
node = Node(label=l, properties={"val": idx})
redis_graph.add_node(node)
nodes[l] = node
edge = Edge(nodes["Src"], "R", nodes["Dest"])
redis_graph.add_edge(edge)
edge = Edge(nodes["Src2"], "R", nodes["Dest"])
redis_graph.add_edge(edge)
redis_graph.commit()
# Delete a node.
query = """MATCH (n:Src2) DELETE n"""
actual_result = redis_graph.query(query)
self.env.assertEquals(actual_result.nodes_deleted, 1)
self.env.assertEquals(actual_result.relationships_deleted, 1)
query = """MATCH (n1:Src)-[*]->(n2:Dest) RETURN COUNT(*)"""
actual_result = redis_graph.query(query)
expected_result = [[1]]
self.env.assertEquals(actual_result.result_set, expected_result)
# Perform the same traversal, this time traveling from destination to source.
query = """MATCH (n1:Src)-[*]->(n2:Dest {val: 0}) RETURN COUNT(*)"""
actual_result = redis_graph.query(query)
expected_result = [[1]]
self.env.assertEquals(actual_result.result_set, expected_result)
class testQueryTimeout(FlowTestsBase):
def __init__(self):
self.env = Env(decodeResponses=True)
# skip test if we're running under Valgrind
if self.env.envRunner.debugger is not None or os.getenv('COV') == '1':
self.env.skip() # queries will be much slower under Valgrind
global redis_con
global redis_graph
redis_con = self.env.getConnection()
redis_graph = Graph("timeout", redis_con)
def test01_read_query_timeout(self):
query = "UNWIND range(0,100000) AS x WITH x AS x WHERE x = 10000 RETURN x"
try:
# The query is expected to timeout
redis_graph.query(query, timeout=1)
assert (False)
except ResponseError as error:
self.env.assertContains("Query timed out", str(error))
try:
# The query is expected to succeed
redis_graph.query(query, timeout=200)
except:
assert (False)
def test02_configured_timeout(self):
# Verify that the module-level timeout is set to the default of 0
response = redis_con.execute_command("GRAPH.CONFIG GET timeout")
self.env.assertEquals(response[1], 0)
# Set a default timeout of 1 millisecond
redis_con.execute_command("GRAPH.CONFIG SET timeout 1")
response = redis_con.execute_command("GRAPH.CONFIG GET timeout")
self.env.assertEquals(response[1], 1)
# Validate that a read query times out
query = "UNWIND range(0,100000) AS x WITH x AS x WHERE x = 10000 RETURN x"
try:
redis_graph.query(query)
assert (False)
except ResponseError as error:
self.env.assertContains("Query timed out", str(error))
def test03_write_query_ignore_timeout(self):
#----------------------------------------------------------------------
# verify that the timeout argument is ignored by write queries
#----------------------------------------------------------------------
write_queries = [
# create query
"UNWIND range(0, 10000) AS x CREATE (a:M)",
# update query
"MATCH (a:M) SET a.v = 2",
# delete query
"MATCH (a:M) DELETE a"
]
# queries should complete successfully
for q in write_queries:
try:
result = redis_graph.query(q, timeout=1)
# the query should have taken longer than the timeout value
self.env.assertGreater(result.run_time_ms, 2)
except ResponseError:
assert (False)
#----------------------------------------------------------------------
# index creation should ignore timeouts
#----------------------------------------------------------------------
query = "UNWIND range (0, 100000) AS x CREATE (:M {v:x})"
redis_graph.query(query)
# create index
query = "CREATE INDEX ON :M(v)"
try:
# the query should complete successfully
result = redis_graph.query(query, timeout=1)
self.env.assertEquals(result.indices_created, 1)
except ResponseError:
assert (False)
#----------------------------------------------------------------------
# index deletion should ignore timeouts
#----------------------------------------------------------------------
query = "DROP INDEX ON :M(v)"
try:
# the query should complete successfully
result = redis_graph.query(query, timeout=1)
self.env.assertEquals(result.indices_deleted, 1)
except ResponseError:
assert (False)
def test04_timeout_index_scan(self):
# construct a graph and create multiple indices
query = """UNWIND range(0, 100000) AS x CREATE (p:Person {age: x%90, height: x%200, weight: x%80})"""
redis_graph.query(query)
query = """CREATE INDEX ON :Person(age, height, weight)"""
redis_graph.query(query)
queries = [
# full scan
"MATCH (a) RETURN a",
# ID scan
"MATCH (a) WHERE ID(a) > 20 RETURN a",
# label scan
"MATCH (a:Person) RETURN a",
# single index scan
"MATCH (a:Person) WHERE a.age > 40 RETURN a",
# index scan + full scan
"MATCH (a:Person), (b) WHERE a.age > 40 RETURN a, b",
# index scan + ID scan
"MATCH (a:Person), (b) WHERE a.age > 40 AND ID(b) > 20 RETURN a, b",
# index scan + label scan
"MATCH (a:Person), (b:Person) WHERE a.age > 40 RETURN a, b",
# multi full and index scans
"MATCH (a:Person), (b:Person), (c), (d) WHERE a.age > 40 AND b.height < 150 RETURN a,b,c,d",
# multi ID and index scans
"MATCH (a:Person), (b:Person), (c:Person), (d) WHERE a.age > 40 AND b.height < 150 AND ID(c) > 20 AND ID(d) > 30 RETURN a,b,c,d",
# multi label and index scans
"MATCH (a:Person), (b:Person), (c:Person), (d:Person) WHERE a.age > 40 AND b.height < 150 RETURN a,b,c,d",
# multi index scans
"MATCH (a:Person), (b:Person), (c:Person) WHERE a.age > 40 AND b.height < 150 AND c.weight = 50 RETURN a,b,c"
]
for q in queries:
try:
# query is expected to timeout
redis_graph.query(q, timeout=2)
assert (False)
except ResponseError as error:
self.env.assertContains("Query timed out", str(error))
# validate that server didn't crash
redis_con.ping()
# rerun each query with timeout and limit
# expecting queries to run to completion
for q in queries:
q += " LIMIT 2"
redis_graph.query(q, timeout=10)
# validate that server didn't crash
redis_con.ping()
class testEdgeIndexUpdatesFlow(FlowTestsBase):
def __init__(self):
self.env = Env(decodeResponses=True)
global redis_graph
redis_con = self.env.getConnection()
redis_graph = Graph(GRAPH_ID, redis_con)
self.populate_graph()
self.build_indices()
def new_node(self):
return Node(label = labels[node_ctr % 2],
properties = {'unique': node_ctr,
'group': random.choice(groups),
'doubleval': round(random.uniform(-1, 1), 2),
'intval': random.randint(1, 10000),
'stringval': ''.join(random.choice(string.ascii_lowercase) for x in range(6))})
def new_edge(self, from_node, to_node):
return Edge(from_node, types[edge_ctr % 2], to_node,
properties={'unique': edge_ctr,
'group': random.choice(groups),
'doubleval': round(random.uniform(-1, 1), 2),
'intval': random.randint(1, 10000),
'stringval': ''.join(random.choice(string.ascii_lowercase) for x in range(6))})
def populate_graph(self):
global node_ctr
global edge_ctr
for i in range(1000):
from_node = self.new_node()
node_ctr += 1
redis_graph.add_node(from_node)
to_node = self.new_node()
node_ctr += 1
redis_graph.add_node(to_node)
edge = self.new_edge(from_node, to_node)
edge_ctr += 1
redis_graph.add_edge(edge)
redis_graph.commit()
def build_indices(self):
for field in fields:
redis_graph.query("CREATE INDEX FOR ()-[r:type_a]-() ON (r.%s)" % (field))
redis_graph.query("CREATE INDEX FOR ()-[r:type_b]-() ON (r.%s)" % (field))
# Validate that all properties are indexed
def validate_indexed(self):
for field in fields:
resp = redis_graph.execution_plan("""MATCH ()-[a:type_a]->() WHERE a.%s > 0 RETURN a""" % (field))
self.env.assertIn('Edge By Index Scan', resp)
# So long as 'unique' is not modified, label_a.unique will always be even and label_b.unique will always be odd
def validate_unique(self):
result = redis_graph.query("MATCH ()-[r:type_a]->() RETURN r.unique")
# Remove the header
result.result_set.pop(0)
for val in result.result_set:
self.env.assertEquals(int(float(val[0])) % 2, 0)
result = redis_graph.query("MATCH ()-[r:type_b]->() RETURN r.unique")
# Remove the header
result.result_set.pop(0)
for val in result.result_set:
self.env.assertEquals(int(float(val[0])) % 2, 1)
# The index scan ought to return identical results to a label scan over the same range of values.
def validate_doubleval(self):
for type in types:
resp = redis_graph.execution_plan("""MATCH ()-[a:%s]->() WHERE a.doubleval < 100 RETURN a.doubleval ORDER BY a.doubleval""" % (type))
self.env.assertIn('Edge By Index Scan', resp)
indexed_result = redis_graph.query("""MATCH ()-[a:%s]->() WHERE a.doubleval < 100 RETURN a.doubleval ORDER BY a.doubleval""" % (type))
scan_result = redis_graph.query("""MATCH ()-[a:%s]->() RETURN a.doubleval ORDER BY a.doubleval""" % (type))
self.env.assertEqual(len(indexed_result.result_set), len(scan_result.result_set))
# Collect any elements between the two result sets that fail a string comparison
# so that we may compare them as doubles (specifically, -0 and 0 should be considered equal)
differences = [[i[0], j[0]] for i, j in zip(indexed_result.result_set, scan_result.result_set) if i != j]
for pair in differences:
self.env.assertEqual(float(pair[0]), float(pair[1]))
# The intval property can be assessed similar to doubleval, but the result sets should be identical
def validate_intval(self):
for type in types:
resp = redis_graph.execution_plan("""MATCH ()-[a:%s]->() WHERE a.intval > 0 RETURN a.intval ORDER BY a.intval""" % (type))
self.env.assertIn('Edge By Index Scan', resp)
indexed_result = redis_graph.query("""MATCH ()-[a:%s]->() WHERE a.intval > 0 RETURN a.intval ORDER BY a.intval""" % (type))
scan_result = redis_graph.query("""MATCH ()-[a:%s]->() RETURN a.intval ORDER BY a.intval""" % (type))
self.env.assertEqual(indexed_result.result_set, scan_result.result_set)
# Validate a series of premises to ensure that the graph has not been modified unexpectedly
def validate_state(self):
self.validate_unique()
self.validate_indexed()
self.validate_doubleval()
self.validate_intval()
# Modify a property, triggering updates to all edges in two indices
def test01_full_property_update(self):
result = redis_graph.query("MATCH ()-[a]->() SET a.doubleval = a.doubleval + %f" % (round(random.uniform(-1, 1), 2)))
self.env.assertEquals(result.properties_set, 1000)
# Verify that index scans still function and return correctly
self.validate_state()
# Modify a property, triggering updates to a subset of edges in two indices
def test02_partial_property_update(self):
redis_graph.query("MATCH ()-[a]->() WHERE a.doubleval > 0 SET a.doubleval = a.doubleval + %f" % (round(random.uniform(-1, 1), 2)))
# Verify that index scans still function and return correctly
self.validate_state()
# Add 100 randomized edges and validate indices
def test03_edge_creation(self):
# Reset nodes in the Graph object so that we won't double-commit the originals
redis_graph.nodes = {}
redis_graph.edges = []
global node_ctr
global edge_ctr
for i in range(100):
from_node = self.new_node()
redis_graph.add_node(from_node)
node_ctr += 1
to_node = self.new_node()
redis_graph.add_node(to_node)
node_ctr += 1
edge = self.new_edge(from_node, to_node)
redis_graph.add_edge(edge)
edge_ctr += 1
redis_graph.commit()
self.validate_state()
# Delete every other edge in first 100 and validate indices
def test04_edge_deletion(self):
# Reset nodes in the Graph object so that we won't double-commit the originals
redis_graph.nodes = {}
redis_graph.edges = []
global node_ctr
global edge_ctr
# Delete edges one at a time
for i in range(0, 100, 2):
result = redis_graph.query("MATCH ()-[a]->() WHERE ID(a) = %d DELETE a" % (i))
self.env.assertEquals(result.relationships_deleted, 1)
edge_ctr -= 1
self.validate_state()
# Delete all edges matching a filter
result = redis_graph.query("MATCH ()-[a:type_a]->() WHERE a.group = 'Group A' DELETE a")
self.env.assertGreater(result.relationships_deleted, 0)
self.validate_state()
def test05_unindexed_property_update(self):
# Add an unindexed property to all edges.
redis_graph.query("MATCH ()-[a]->() SET a.unindexed = 'unindexed'")
# Retrieve a single edge
result = redis_graph.query("MATCH ()-[a]->() RETURN a.unique LIMIT 1")
unique_prop = result.result_set[0][0]
query = """MATCH ()-[a {unique: %s }]->() SET a.unindexed = 5, a.unique = %s RETURN a.unindexed, a.unique""" % (unique_prop, unique_prop)
result = redis_graph.query(query)
expected_result = [[5, unique_prop]]
self.env.assertEquals(result.result_set, expected_result)
self.env.assertEquals(result.properties_set, 1)
# Validate that after deleting an indexed property, that property can no longer be found in the index.
def test06_remove_indexed_prop(self):
# Create a new edge with a single indexed property
query = """CREATE ()-[:NEW {v: 5}]->()"""
result = redis_graph.query(query)
self.env.assertEquals(result.properties_set, 1)
redis_graph.query("CREATE INDEX FOR ()-[r:NEW]-() ON (r.v)")
# Delete the entity's property
query = """MATCH ()-[a:NEW {v: 5}]->() SET a.v = NULL"""
result = redis_graph.query(query)
self.env.assertEquals(result.properties_set, 1)
# Query the index for the entity
query = """MATCH ()-[a:NEW {v: 5}]->() RETURN a"""
plan = redis_graph.execution_plan(query)
self.env.assertIn("Edge By Index Scan", plan)
result = redis_graph.query(query)
# No entities should be returned
expected_result = []
self.env.assertEquals(result.result_set, expected_result)
class testIndexUpdatesFlow(FlowTestsBase):
def __init__(self):
self.env = Env(decodeResponses=True)
global redis_graph
redis_con = self.env.getConnection()
redis_graph = Graph(GRAPH_ID, redis_con)
self.populate_graph()
self.build_indices()
def new_node(self):
return Node(label=labels[node_ctr % 2],
properties={
'unique':
node_ctr,
'group':
random.choice(groups),
'doubleval':
round(random.uniform(-1, 1), 2),
'intval':
random.randint(1, 10000),
'stringval':
''.join(
random.choice(string.ascii_lowercase)
for x in range(6))
})
def populate_graph(self):
global node_ctr
for i in range(1000):
node = self.new_node()
redis_graph.add_node(node)
node_ctr += 1
redis_graph.commit()
def build_indices(self):
for field in fields:
redis_graph.query("CREATE INDEX ON :label_a(%s)" % (field))
redis_graph.query("CREATE INDEX ON :label_b(%s)" % (field))
# Validate that all properties are indexed
def validate_indexed(self):
for field in fields:
resp = redis_graph.execution_plan(
"""MATCH (a:label_a) WHERE a.%s > 0 RETURN a""" % (field))
self.env.assertIn('Node By Index Scan', resp)
# So long as 'unique' is not modified, label_a.unique will always be even and label_b.unique will always be odd
def validate_unique(self):
result = redis_graph.query("MATCH (a:label_a) RETURN a.unique")
# Remove the header
result.result_set.pop(0)
for val in result.result_set:
self.env.assertEquals(int(float(val[0])) % 2, 0)
result = redis_graph.query("MATCH (b:label_b) RETURN b.unique")
# Remove the header
result.result_set.pop(0)
for val in result.result_set:
self.env.assertEquals(int(float(val[0])) % 2, 1)
# The index scan ought to return identical results to a label scan over the same range of values.
def validate_doubleval(self):
for label in labels:
resp = redis_graph.execution_plan(
"""MATCH (a:%s) WHERE a.doubleval < 100 RETURN a.doubleval ORDER BY a.doubleval"""
% (label))
self.env.assertIn('Node By Index Scan', resp)
indexed_result = redis_graph.query(
"""MATCH (a:%s) WHERE a.doubleval < 100 RETURN a.doubleval ORDER BY a.doubleval"""
% (label))
scan_result = redis_graph.query(
"""MATCH (a:%s) RETURN a.doubleval ORDER BY a.doubleval""" %
(label))
self.env.assertEqual(len(indexed_result.result_set),
len(scan_result.result_set))
# Collect any elements between the two result sets that fail a string comparison
# so that we may compare them as doubles (specifically, -0 and 0 should be considered equal)
differences = [[i[0], j[0]] for i, j in zip(
indexed_result.result_set, scan_result.result_set) if i != j]
for pair in differences:
self.env.assertEqual(float(pair[0]), float(pair[1]))
# The intval property can be assessed similar to doubleval, but the result sets should be identical
def validate_intval(self):
for label in labels:
resp = redis_graph.execution_plan(
"""MATCH (a:%s) WHERE a.intval > 0 RETURN a.intval ORDER BY a.intval"""
% (label))
self.env.assertIn('Node By Index Scan', resp)
indexed_result = redis_graph.query(
"""MATCH (a:%s) WHERE a.intval > 0 RETURN a.intval ORDER BY a.intval"""
% (label))
scan_result = redis_graph.query(
"""MATCH (a:%s) RETURN a.intval ORDER BY a.intval""" % (label))
self.env.assertEqual(indexed_result.result_set,
scan_result.result_set)
# Validate a series of premises to ensure that the graph has not been modified unexpectedly
def validate_state(self):
self.validate_unique()
self.validate_indexed()
self.validate_doubleval()
self.validate_intval()
# Modify a property, triggering updates to all nodes in two indices
def test01_full_property_update(self):
result = redis_graph.query(
"MATCH (a) SET a.doubleval = a.doubleval + %f" %
(round(random.uniform(-1, 1), 2)))
self.env.assertEquals(result.properties_set, 1000)
# Verify that index scans still function and return correctly
self.validate_state()
# Modify a property, triggering updates to a subset of nodes in two indices
def test02_partial_property_update(self):
redis_graph.query(
"MATCH (a) WHERE a.doubleval > 0 SET a.doubleval = a.doubleval + %f"
% (round(random.uniform(-1, 1), 2)))
# Verify that index scans still function and return correctly
self.validate_state()
# Add 100 randomized nodes and validate indices
def test03_node_creation(self):
# Reset nodes in the Graph object so that we won't double-commit the originals
redis_graph.nodes = {}
global node_ctr
for i in range(100):
node = self.new_node()
redis_graph.add_node(node)
node_ctr += 1
redis_graph.commit()
self.validate_state()
# Delete every other node in first 100 and validate indices
def test04_node_deletion(self):
# Reset nodes in the Graph object so that we won't double-commit the originals
redis_graph.nodes = {}
global node_ctr
# Delete nodes one at a time
for i in range(0, 100, 2):
result = redis_graph.query("MATCH (a) WHERE ID(a) = %d DELETE a" %
(i))
self.env.assertEquals(result.nodes_deleted, 1)
node_ctr -= 1
self.validate_state()
# Delete all nodes matching a filter
result = redis_graph.query(
"MATCH (a:label_a) WHERE a.group = 'Group A' DELETE a")
self.env.assertGreater(result.nodes_deleted, 0)
self.validate_state()
def test05_unindexed_property_update(self):
# Add an unindexed property to all nodes.
redis_graph.query("MATCH (a) SET a.unindexed = 'unindexed'")
# Retrieve a single node
result = redis_graph.query("MATCH (a) RETURN a.unique LIMIT 1")
unique_prop = result.result_set[0][0]
query = """MATCH (a {unique: %s }) SET a.unindexed = 5, a.unique = %s RETURN a.unindexed, a.unique""" % (
unique_prop, unique_prop)
result = redis_graph.query(query)
expected_result = [[5, unique_prop]]
self.env.assertEquals(result.result_set, expected_result)
self.env.assertEquals(result.properties_set, 1)
# Validate that after deleting an indexed property, that property can no longer be found in the index.
def test06_remove_indexed_prop(self):
# Create a new node with a single indexed property
query = """CREATE (:NEW {v: 5})"""
result = redis_graph.query(query)
self.env.assertEquals(result.properties_set, 1)
self.env.assertEquals(result.labels_added, 1)
redis_graph.query("CREATE INDEX ON :NEW(v)")
# Delete the entity's property
query = """MATCH (a:NEW {v: 5}) SET a.v = NULL"""
result = redis_graph.query(query)
self.env.assertEquals(result.properties_set, 1)
# Query the index for the entity
query = """MATCH (a:NEW {v: 5}) RETURN a"""
plan = redis_graph.execution_plan(query)
self.env.assertIn("Node By Index Scan", plan)
result = redis_graph.query(query)
# No entities should be returned
expected_result = []
self.env.assertEquals(result.result_set, expected_result)
# Validate that when a label has both exact-match and full-text indexes
# on different properties, an update operation checks all indexes to
# determine whether they must be updated.
# This is necessary because either one of the indexes may not track the
# property being updated, but that does not guarantee that the other
# index does not track the property.
def test07_update_property_only_on_fulltext_index(self):
# Remove the exact-match index on a property
redis_graph.redis_con.execute_command("GRAPH.QUERY", GRAPH_ID,
"DROP INDEX ON :label_a(group)")
# Add a full-text index on the property
redis_graph.query(
"CALL db.idx.fulltext.createNodeIndex('label_a', 'group')")
# Modify the values of the property
result = redis_graph.query(
"MATCH (a:label_a) WHERE a.group = 'Group C' SET a.group = 'Group NEW'"
)
modified_count = result.properties_set
self.env.assertGreater(modified_count, 0)
# Validate that the full-text index reflects the update
result = redis_graph.query(
"CALL db.idx.fulltext.queryNodes('label_a', 'Group NEW')")
self.env.assertEquals(len(result.result_set), modified_count)
# Validate that the previous value has been removed
result = redis_graph.query(
"CALL db.idx.fulltext.queryNodes('label_a', 'Group C')")
self.env.assertEquals(len(result.result_set), 0)
class TestAggregate():
def __init__(self):
self.env = Env()
add_values(self.env)
def testGroupBy(self):
cmd = ['ft.aggregate', 'games', '*',
'GROUPBY', '1', '@brand',
'REDUCE', 'count', '0', 'AS', 'count',
'SORTBY', 2, '@count', 'desc',
'LIMIT', '0', '5'
]
res = self.env.cmd(*cmd)
self.env.assertIsNotNone(res)
self.env.assertEqual([292L, ['brand', '', 'count', '1518'], ['brand', 'mad catz', 'count', '43'],
['brand', 'generic', 'count', '40'], ['brand', 'steelseries', 'count', '37'],
['brand', 'logitech', 'count', '35']], res)
def testMinMax(self):
cmd = ['ft.aggregate', 'games', 'sony',
'GROUPBY', '1', '@brand',
'REDUCE', 'count', '0',
'REDUCE', 'min', '1', '@price', 'as', 'minPrice',
'SORTBY', '2', '@minPrice', 'DESC']
res = self.env.cmd(*cmd)
self.env.assertIsNotNone(res)
row = to_dict(res[1])
self.env.assertEqual(88, int(float(row['minPrice'])))
cmd = ['ft.aggregate', 'games', 'sony',
'GROUPBY', '1', '@brand',
'REDUCE', 'count', '0',
'REDUCE', 'max', '1', '@price', 'as', 'maxPrice',
'SORTBY', '2', '@maxPrice', 'DESC']
res = self.env.cmd(*cmd)
row = to_dict(res[1])
self.env.assertEqual(695, int(float(row['maxPrice'])))
def testAvg(self):
cmd = ['ft.aggregate', 'games', 'sony',
'GROUPBY', '1', '@brand',
'REDUCE', 'avg', '1', '@price', 'AS', 'avg_price',
'REDUCE', 'count', '0',
'SORTBY', '2', '@avg_price', 'DESC']
res = self.env.cmd(*cmd)
self.env.assertIsNotNone(res)
self.env.assertEqual(26, res[0])
# Ensure the formatting actually exists
first_row = to_dict(res[1])
self.env.assertEqual(109, int(float(first_row['avg_price'])))
for row in res[1:]:
row = to_dict(row)
self.env.assertIn('avg_price', row)
# Test aliasing
cmd = ['FT.AGGREGATE', 'games', 'sony', 'GROUPBY', '1', '@brand',
'REDUCE', 'avg', '1', '@price', 'AS', 'avgPrice']
res = self.env.cmd(*cmd)
first_row = to_dict(res[1])
self.env.assertEqual(17, int(float(first_row['avgPrice'])))
def testCountDistinct(self):
cmd = ['FT.AGGREGATE', 'games', '*',
'GROUPBY', '1', '@brand',
'REDUCE', 'COUNT_DISTINCT', '1', '@title', 'AS', 'count_distinct(title)',
'REDUCE', 'COUNT', '0'
]
res = self.env.cmd(*cmd)[1:]
# print res
row = to_dict(res[0])
self.env.assertEqual(1484, int(row['count_distinct(title)']))
cmd = ['FT.AGGREGATE', 'games', '*',
'GROUPBY', '1', '@brand',
'REDUCE', 'COUNT_DISTINCTISH', '1', '@title', 'AS', 'count_distinctish(title)',
'REDUCE', 'COUNT', '0'
]
res = self.env.cmd(*cmd)[1:]
# print res
row = to_dict(res[0])
self.env.assertEqual(1461, int(row['count_distinctish(title)']))
def testQuantile(self):
cmd = ['FT.AGGREGATE', 'games', '*',
'GROUPBY', '1', '@brand',
'REDUCE', 'QUANTILE', '2', '@price', '0.50', 'AS', 'q50',
'REDUCE', 'QUANTILE', '2', '@price', '0.90', 'AS', 'q90',
'REDUCE', 'QUANTILE', '2', '@price', '0.95', 'AS', 'q95',
'REDUCE', 'AVG', '1', '@price',
'REDUCE', 'COUNT', '0', 'AS', 'rowcount',
'SORTBY', '2', '@rowcount', 'DESC', 'MAX', '1']
res = self.env.cmd(*cmd)
row = to_dict(res[1])
# TODO: Better samples
self.env.assertAlmostEqual(14.99, float(row['q50']), delta=3)
self.env.assertAlmostEqual(70, float(row['q90']), delta=50)
# This tests the 95th percentile, which is error prone because
# so few samples actually exist. I'm disabling it for now so that
# there is no breakage in CI
# self.env.assertAlmostEqual(110, (float(row['q95'])), delta=50)
def testStdDev(self):
cmd = ['FT.AGGREGATE', 'games', '*',
'GROUPBY', '1', '@brand',
'REDUCE', 'STDDEV', '1', '@price', 'AS', 'stddev(price)',
'REDUCE', 'AVG', '1', '@price', 'AS', 'avgPrice',
'REDUCE', 'QUANTILE', '2', '@price', '0.50', 'AS', 'q50Price',
'REDUCE', 'COUNT', '0', 'AS', 'rowcount',
'SORTBY', '2', '@rowcount', 'DESC',
'LIMIT', '0', '10']
res = self.env.cmd(*cmd)
row = to_dict(res[1])
self.env.assertTrue(10 <= int(
float(row['q50Price'])) <= 20)
self.env.assertAlmostEqual(53, int(float(row['stddev(price)'])), delta=50)
self.env.assertEqual(29, int(float(row['avgPrice'])))
def testParseTime(self):
cmd = ['FT.AGGREGATE', 'games', '*',
'GROUPBY', '1', '@brand',
'REDUCE', 'COUNT', '0', 'AS', 'count',
'APPLY', 'timefmt(1517417144)', 'AS', 'dt',
'APPLY', 'parse_time("%FT%TZ", @dt)', 'as', 'parsed_dt',
'LIMIT', '0', '1']
res = self.env.cmd(*cmd)
self.env.assertEqual(['brand', '', 'count', '1518', 'dt',
'2018-01-31T16:45:44Z', 'parsed_dt', '1517417144'], res[1])
def testRandomSample(self):
cmd = ['FT.AGGREGATE', 'games', '*', 'GROUPBY', '1', '@brand',
'REDUCE', 'COUNT', '0', 'AS', 'num',
'REDUCE', 'RANDOM_SAMPLE', '2', '@price', '10',
'SORTBY', '2', '@num', 'DESC', 'MAX', '10']
for row in self.env.cmd(*cmd)[1:]:
self.env.assertIsInstance(row[5], list)
self.env.assertGreater(len(row[5]), 0)
self.env.assertGreaterEqual(row[3], len(row[5]))
self.env.assertLessEqual(len(row[5]), 10)
def testTimeFunctions(self):
cmd = ['FT.AGGREGATE', 'games', '*',
'APPLY', '1517417144', 'AS', 'dt',
'APPLY', 'timefmt(@dt)', 'AS', 'timefmt',
'APPLY', 'day(@dt)', 'AS', 'day',
'APPLY', 'hour(@dt)', 'AS', 'hour',
'APPLY', 'minute(@dt)', 'AS', 'minute',
'APPLY', 'month(@dt)', 'AS', 'month',
'APPLY', 'dayofweek(@dt)', 'AS', 'dayofweek',
'APPLY', 'dayofmonth(@dt)', 'AS', 'dayofmonth',
'APPLY', 'dayofyear(@dt)', 'AS', 'dayofyear',
'APPLY', 'year(@dt)', 'AS', 'year',
'LIMIT', '0', '1']
res = self.env.cmd(*cmd)
self.env.assertListEqual([1L, ['dt', '1517417144', 'timefmt', '2018-01-31T16:45:44Z', 'day', '1517356800', 'hour', '1517414400',
'minute', '1517417100', 'month', '1514764800', 'dayofweek', '3', 'dayofmonth', '31', 'dayofyear', '30', 'year', '2018']], res)
def testStringFormat(self):
cmd = ['FT.AGGREGATE', 'games', '@brand:sony',
'GROUPBY', '2', '@title', '@brand',
'REDUCE', 'COUNT', '0',
'REDUCE', 'MAX', '1', '@price', 'AS', 'price',
'APPLY', 'format("%s|%s|%s|%s", @title, @brand, "Mark", @price)', 'as', 'titleBrand',
'LIMIT', '0', '10']
res = self.env.cmd(*cmd)
for row in res[1:]:
row = to_dict(row)
expected = '%s|%s|%s|%g' % (
row['title'], row['brand'], 'Mark', float(row['price']))
self.env.assertEqual(expected, row['titleBrand'])
def testSum(self):
cmd = ['ft.aggregate', 'games', '*',
'GROUPBY', '1', '@brand',
'REDUCE', 'count', '0', 'AS', 'count',
'REDUCE', 'sum', 1, '@price', 'AS', 'sum(price)',
'SORTBY', 2, '@sum(price)', 'desc',
'LIMIT', '0', '5'
]
res = self.env.cmd(*cmd)
self.env.assertEqual([292L, ['brand', '', 'count', '1518', 'sum(price)', '44780.69'],
['brand', 'mad catz', 'count',
'43', 'sum(price)', '3973.48'],
['brand', 'razer', 'count', '26',
'sum(price)', '2558.58'],
['brand', 'logitech', 'count',
'35', 'sum(price)', '2329.21'],
['brand', 'steelseries', 'count', '37', 'sum(price)', '1851.12']], res)
def testFilter(self):
cmd = ['ft.aggregate', 'games', '*',
'GROUPBY', '1', '@brand',
'REDUCE', 'count', '0', 'AS', 'count',
'FILTER', '@count > 5'
]
res = self.env.cmd(*cmd)
for row in res[1:]:
row = to_dict(row)
self.env.assertGreater(int(row['count']), 5)
cmd = ['ft.aggregate', 'games', '*',
'GROUPBY', '1', '@brand',
'REDUCE', 'count', '0', 'AS', 'count',
'FILTER', '@count < 5',
'FILTER', '@count > 2 && @brand != ""'
]
res = self.env.cmd(*cmd)
for row in res[1:]:
row = to_dict(row)
self.env.assertLess(int(row['count']), 5)
self.env.assertGreater(int(row['count']), 2)
def testToList(self):
cmd = ['ft.aggregate', 'games', '*',
'GROUPBY', '1', '@brand',
'REDUCE', 'count_distinct', '1', '@price', 'as', 'count',
'REDUCE', 'tolist', 1, '@price', 'as', 'prices',
'SORTBY', 2, '@count', 'desc',
'LIMIT', '0', '5'
]
res = self.env.cmd(*cmd)
for row in res[1:]:
row = to_dict(row)
self.env.assertEqual(int(row['count']), len(row['prices']))
def testSortBy(self):
res = self.env.cmd('ft.aggregate', 'games', '*', 'GROUPBY', '1', '@brand',
'REDUCE', 'sum', 1, '@price', 'as', 'price',
'SORTBY', 2, '@price', 'desc',
'LIMIT', '0', '2')
self.env.assertListEqual([292L, ['brand', '', 'price', '44780.69'], [
'brand', 'mad catz', 'price', '3973.48']], res)
res = self.env.cmd('ft.aggregate', 'games', '*', 'GROUPBY', '1', '@brand',
'REDUCE', 'sum', 1, '@price', 'as', 'price',
'SORTBY', 2, '@price', 'asc',
'LIMIT', '0', '2')
self.env.assertListEqual([292L, ['brand', 'myiico', 'price', '0.23'], [
'brand', 'crystal dynamics', 'price', '0.25']], res)
# Test MAX with limit higher than it
res = self.env.cmd('ft.aggregate', 'games', '*', 'GROUPBY', '1', '@brand',
'REDUCE', 'sum', 1, '@price', 'as', 'price',
'SORTBY', 2, '@price', 'asc', 'MAX', 2)
self.env.assertListEqual([292L, ['brand', 'myiico', 'price', '0.23'], [
'brand', 'crystal dynamics', 'price', '0.25']], res)
# Test Sorting by multiple properties
res = self.env.cmd('ft.aggregate', 'games', '*', 'GROUPBY', '1', '@brand',
'REDUCE', 'sum', 1, '@price', 'as', 'price',
'APPLY', '(@price % 10)', 'AS', 'price',
'SORTBY', 4, '@price', 'asc', '@brand', 'desc', 'MAX', 10,
)
self.env.assertListEqual([292L, ['brand', 'zps', 'price', '0'], ['brand', 'zalman', 'price', '0'], ['brand', 'yoozoo', 'price', '0'], ['brand', 'white label', 'price', '0'], ['brand', 'stinky', 'price', '0'], [
'brand', 'polaroid', 'price', '0'], ['brand', 'plantronics', 'price', '0'], ['brand', 'ozone', 'price', '0'], ['brand', 'oooo', 'price', '0'], ['brand', 'neon', 'price', '0']], res)
def testExpressions(self):
pass
def testNoGroup(self):
res = self.env.cmd('ft.aggregate', 'games', '*', 'LOAD', '2', '@brand', '@price',
'APPLY', 'floor(sqrt(@price)) % 10', 'AS', 'price',
'SORTBY', 4, '@price', 'desc', '@brand', 'desc', 'MAX', 5,
)
exp = [2265L,
['brand', 'xbox', 'price', '9'],
['brand', 'turtle beach', 'price', '9'],
['brand', 'trust', 'price', '9'],
['brand', 'steelseries', 'price', '9'],
['brand', 'speedlink', 'price', '9']]
# exp = [2265L, ['brand', 'Xbox', 'price', '9'], ['brand', 'Turtle Beach', 'price', '9'], [
# 'brand', 'Trust', 'price', '9'], ['brand', 'SteelSeries', 'price', '9'], ['brand', 'Speedlink', 'price', '9']]
self.env.assertListEqual(exp[1], res[1])
def testLoad(self):
res = self.env.cmd('ft.aggregate', 'games', '*',
'LOAD', '3', '@brand', '@price', '@nonexist',
'SORTBY', 2, '@price', 'DESC', 'MAX', 2)
exp = [3L, ['brand', '', 'price', '759.12'], ['brand', 'Sony', 'price', '695.8']]
self.env.assertEqual(exp[1], res[1])
def testSplit(self):
res = self.env.cmd('ft.aggregate', 'games', '*', 'APPLY', 'split("hello world, foo,,,bar,", ",", " ")', 'AS', 'strs',
'APPLY', 'split("hello world, foo,,,bar,", " ", ",")', 'AS', 'strs2',
'APPLY', 'split("hello world, foo,,,bar,", "", "")', 'AS', 'strs3',
'APPLY', 'split("hello world, foo,,,bar,")', 'AS', 'strs4',
'APPLY', 'split("hello world, foo,,,bar,",",")', 'AS', 'strs5',
'APPLY', 'split("")', 'AS', 'empty',
'LIMIT', '0', '1'
)
# print "Got {} results".format(len(res))
# return
# pprint.pprint(res)
self.env.assertListEqual([1L, ['strs', ['hello world', 'foo', 'bar'],
'strs2', ['hello', 'world', 'foo,,,bar'],
'strs3', ['hello world, foo,,,bar,'],
'strs4', ['hello world', 'foo', 'bar'],
'strs5', ['hello world', 'foo', 'bar'],
'empty', []]], res)
def testFirstValue(self):
res = self.env.cmd('ft.aggregate', 'games', '@brand:(sony|matias|beyerdynamic|(mad catz))',
'GROUPBY', 1, '@brand',
'REDUCE', 'FIRST_VALUE', 4, '@title', 'BY', '@price', 'DESC', 'AS', 'top_item',
'REDUCE', 'FIRST_VALUE', 4, '@price', 'BY', '@price', 'DESC', 'AS', 'top_price',
'REDUCE', 'FIRST_VALUE', 4, '@title', 'BY', '@price', 'ASC', 'AS', 'bottom_item',
'REDUCE', 'FIRST_VALUE', 4, '@price', 'BY', '@price', 'ASC', 'AS', 'bottom_price',
'SORTBY', 2, '@top_price', 'DESC', 'MAX', 5
)
expected = [4L, ['brand', 'sony', 'top_item', 'sony psp slim & lite 2000 console', 'top_price', '695.8', 'bottom_item', 'sony dlchd20p high speed hdmi cable for playstation 3', 'bottom_price', '5.88'],
['brand', 'matias', 'top_item', 'matias halfkeyboard usb', 'top_price',
'559.99', 'bottom_item', 'matias halfkeyboard usb', 'bottom_price', '559.99'],
['brand', 'beyerdynamic', 'top_item', 'beyerdynamic mmx300 pc gaming premium digital headset with microphone', 'top_price', '359.74',
'bottom_item', 'beyerdynamic headzone pc gaming digital surround sound system with mmx300 digital headset with microphone', 'bottom_price', '0'],
['brand', 'mad catz', 'top_item', 'mad catz s.t.r.i.k.e.7 gaming keyboard', 'top_price', '295.95', 'bottom_item', 'madcatz mov4545 xbox replacement breakaway cable', 'bottom_price', '3.49']]
self.env.assertListEqual(expected, res)
def testLoadAfterGroupBy(self):
with self.env.assertResponseError():
self.env.cmd('ft.aggregate', 'games', '*',
'GROUPBY', 1, '@brand',
'LOAD', 1, '@brand')