6
0
mirror of https://github.com/FirebirdSQL/firebird-qa.git synced 2025-01-22 21:43:06 +01:00
firebird-qa/tests/bugs/core_2987_test.py

169 lines
6.6 KiB
Python
Raw Normal View History

2021-04-26 20:07:00 +02:00
#coding:utf-8
#
# id: bugs.core_2987
# title: Don't send full length of field over the wire when field is null
2021-10-21 19:29:23 +02:00
# decription:
2021-04-26 20:07:00 +02:00
# Checked on: WI-V3.0.0.32484 (SS/SC/CS), LI-T4.0.0.138 (SS)
2021-10-21 19:29:23 +02:00
#
2021-04-26 20:07:00 +02:00
# tracker_id: CORE-2987
# min_versions: ['3.0']
# versions: 3.0
2021-10-21 19:29:23 +02:00
# qmid:
2021-04-26 20:07:00 +02:00
import pytest
from firebird.qa import db_factory, isql_act, Action
# version: 3.0
# resources: None
substitutions_1 = []
init_script_1 = """"""
db_1 = db_factory(sql_dialect=3, init=init_script_1)
test_script_1 = """
2021-10-21 19:29:23 +02:00
-- Measurement showed than on 3.0 (SS/SC/CS) transfer of NULLs is more than 5 (five) times
2021-04-26 20:07:00 +02:00
-- faster than text data with length = 32K. As of 2.5 (SC) than ration is about 1.7 ... 1.8.
-- Test fills up two tables: one with text data and another with only nulls.
-- Then we receive data from these tables via ES/EDS, evaluate elapsed time for both cases
-- and calculate its ratio. This ratio in 3.0 should be not less than 4x.
create or alter procedure sp_test returns(sel_data_ms int, sel_null_ms int) as begin end;
create or alter procedure sp_test as begin end;
recreate global temporary table test1(f01 char(32760)) on commit delete rows;
recreate global temporary table test2(f01 char(32760)) on commit delete rows;
commit;
set term ^;
create or alter procedure sp_fill as
declare c_added_rows int = 2000;
begin
execute statement
'insert into test1 '
|| 'with recursive r as (select 0 i from rdb$database union all select r.i+1 from r where r.i<99) '
|| 'select rpad('''', 32760, uuid_to_char(gen_uuid()) ) from r r1,r r2 rows ' || c_added_rows
on external 'localhost:' || rdb$get_context('SYSTEM','DB_NAME')
2021-10-21 19:29:23 +02:00
as user 'SYSDBA' password 'masterkey'
2021-04-26 20:07:00 +02:00
role 'R001' ------------------------------- this will create new attach #1 that will be used later!
;
execute statement
'insert into test2 '
|| 'with recursive r as (select 0 i from rdb$database union all select r.i+1 from r where r.i<99) '
|| 'select null from r r1,r r2 rows ' || c_added_rows
on external 'localhost:' || rdb$get_context('SYSTEM','DB_NAME')
2021-10-21 19:29:23 +02:00
as user 'SYSDBA' password 'masterkey'
2021-04-26 20:07:00 +02:00
role 'R002' ------------------------------- this will create new attach #2 that will be used later!
;
end
^
create or alter procedure sp_test returns( sel_data_ms int, sel_null_ms int, ratio numeric(10,2) ) as
declare v_f01 type of column test1.f01;
declare t0 timestamp;
declare t1 timestamp;
declare t2 timestamp;
begin
t0='now';
for
execute statement
'select f01 from test1' on external 'localhost:' || rdb$get_context('SYSTEM','DB_NAME')
2021-10-21 19:29:23 +02:00
as user 'SYSDBA' password 'masterkey'
2021-04-26 20:07:00 +02:00
role 'R001' --------------- here we use EXISTING attach #1 (from internal FB connection pool)
into v_f01
do begin end
2021-10-21 19:29:23 +02:00
2021-04-26 20:07:00 +02:00
t1='now';
2021-10-21 19:29:23 +02:00
2021-04-26 20:07:00 +02:00
for
execute statement
'select f01 from test2' on external 'localhost:' || rdb$get_context('SYSTEM','DB_NAME')
2021-10-21 19:29:23 +02:00
as user 'SYSDBA' password 'masterkey'
2021-04-26 20:07:00 +02:00
role 'R002' --------------- here we use EXISTING attach #2 (from internal FB connection pool)
into v_f01
do begin end
2021-10-21 19:29:23 +02:00
2021-04-26 20:07:00 +02:00
t2='now';
2021-10-21 19:29:23 +02:00
2021-04-26 20:07:00 +02:00
sel_data_ms = datediff(millisecond from t0 to t1);
sel_null_ms = datediff(millisecond from t1 to t2);
ratio = coalesce( sel_data_ms * 1.000 / nullif(sel_null_ms,0), 9999999.99);
2021-10-21 19:29:23 +02:00
2021-04-26 20:07:00 +02:00
suspend;
2021-10-21 19:29:23 +02:00
2021-04-26 20:07:00 +02:00
end
^
set term ;^
--commit;
2021-10-21 19:29:23 +02:00
2021-04-26 20:07:00 +02:00
set list on;
2021-10-21 19:29:23 +02:00
2021-04-26 20:07:00 +02:00
--set stat on;
-- must be ~15 sec
execute procedure sp_fill; -- will add data into GTT test1 (in attach #1) and GTT test2 (in attach #2)
--set stat off;
set term ^;
execute block returns( measure_result varchar(40) ) as
declare min_profit_ratio numeric(3,1) = 3.8; -- ##### T H R E S H O L D ####
declare sel_data_ms int;
declare sel_null_ms int;
declare ratio numeric(10,2);
begin
-- These four calls will extract rows from GTT test1 (in attach #1) and from GTT test2 (in attach #2).
-- We have to register statistics only for 2nd...4th measurements due to probable file cache affects.
select ratio, '' from sp_test into ratio, measure_result;
--- do not suspend now --
select sel_data_ms, sel_null_ms, ratio, iif( ratio >= :min_profit_ratio, 'WINS >= '||:min_profit_ratio||'x', 'LOOSES, '||ratio||'x')
from sp_test
into sel_data_ms, sel_null_ms,ratio, measure_result;
suspend;
select sel_data_ms, sel_null_ms, ratio, iif( ratio >= :min_profit_ratio, 'WINS >= '||:min_profit_ratio||'x', 'LOOSES, '||ratio||'x')
from sp_test
into sel_data_ms, sel_null_ms,ratio, measure_result;
suspend;
select sel_data_ms, sel_null_ms, ratio, iif( ratio >= :min_profit_ratio, 'WINS >= '||:min_profit_ratio||'x', 'LOOSES, '||ratio||'x')
from sp_test
into sel_data_ms, sel_null_ms,ratio, measure_result;
suspend;
end
^
set term ;^
commit;
-- ||||||||||||||||||||||||||||
-- ###################################||| FB 4.0+, SS and SC |||##############################
-- ||||||||||||||||||||||||||||
-- If we check SS or SC and ExtConnPoolLifeTime > 0 (config parameter FB 4.0+) then current
-- DB (bugs.core_NNNN.fdb) will be 'captured' by firebird.exe process and fbt_run utility
-- will not able to drop this database at the final point of test.
-- Moreover, DB file will be hold until all activity in firebird.exe completed and AFTER this
-- we have to wait for <ExtConnPoolLifeTime> seconds after it (discussion and small test see
-- in the letter to hvlad and dimitr 13.10.2019 11:10).
-- This means that one need to kill all connections to prevent from exception on cleanup phase:
-- SQLCODE: -901 / lock time-out on wait transaction / object <this_test_DB> is in use
-- #############################################################################################
delete from mon$attachments where mon$attachment_id != current_connection;
2021-10-21 19:29:23 +02:00
commit;
2021-04-26 20:07:00 +02:00
"""
act_1 = isql_act('db_1', test_script_1, substitutions=substitutions_1)
expected_stdout_1 = """
MEASURE_RESULT WINS >= 3.8x
MEASURE_RESULT WINS >= 3.8x
MEASURE_RESULT WINS >= 3.8x
"""
@pytest.mark.version('>=3.0')
2021-04-28 12:42:11 +02:00
def test_1(act_1: Action):
2021-04-26 20:07:00 +02:00
act_1.expected_stdout = expected_stdout_1
act_1.execute()
assert act_1.clean_expected_stdout == act_1.clean_stdout