[MCOL-25] 001 Working Folder UM Join Test Fails Created: 2016-05-03  Updated: 2016-10-03  Resolved: 2016-10-03

Status: Closed
Project: MariaDB ColumnStore
Component/s: PrimProc
Affects Version/s: None
Fix Version/s: 1.0.4

Type: Task Priority: Major
Reporter: Dipti Joshi (Inactive) Assignee: Andrew Hutchings (Inactive)
Resolution: Fixed Votes: 0
Labels: None

Issue Links:
PartOf
includes MCOL-271 Improved support for NULL for Varchar... Closed
includes MCOL-324 concat of long strings truncated at 5... Closed
includes MCOL-325 WEEK() handling needs to match MySQL's Closed
includes MCOL-326 RAND() behaves differently with negat... Closed
includes MCOL-327 time() comparison broken with quoted ... Closed
includes MCOL-328 REVERSE() adds extra character Closed
includes MCOL-329 Functions casting date/datetime 0 to ... Closed
includes MCOL-330 Datetime to int conversion returns st... Closed
includes MCOL-331 INSERT() function misbehaves with our... Closed
includes MCOL-332 MONTHNAME() to int should equal 0 Closed
includes MCOL-333 subtime() off by one second Closed
includes MCOL-334 Bad NULL match in view subquery Closed
is part of MCOL-280 Beta issues Closed
Sprint: 2016-19

 Description   

Working Folder UM Join Test Details:

Working Folder Test scripts that failed:
Ref failed - working_tpch1/aggregation/group_concat.sql
Ref failed - working_tpch1/group/q4.8.9.sql
Local failed - working_tpch1/misc/bug2954.sql
Local failed - working_tpch1/misc/bug3783.sql
Local failed - working_tpch1/misc/bug5267.sql
Local failed - working_tpch1/misc/insert_select.sql
Ref failed - working_tpch1/misc/bug2976.sql
Ref failed - working_tpch1/misc/bug3475.sql
Ref failed - working_tpch1/misc/bug3719.sql
Ref failed - working_tpch1/misc/bug3749.sql
Ref failed - working_tpch1/misc/bug3881.sql
Ref failed - working_tpch1/misc/bug3932.sql
Ref failed - working_tpch1/misc/bug3935.sql
Ref failed - working_tpch1/misc/bug3948.sql
Ref failed - working_tpch1/misc/bug4757.sql
Ref failed - working_tpch1/misc/bug4827.sql
Local failed - working_tpch1/qa_fe_cnxFunctions/CASE1.SM.sql
Local failed - working_tpch1/qa_fe_cnxFunctions/CASE2.SM.sql
Local failed - working_tpch1/qa_fe_cnxFunctions/DAYNAME.DS.sql
Local failed - working_tpch1/qa_fe_cnxFunctions/MONTHNAME.DS.sql
Local failed - working_tpch1/qa_fe_cnxFunctions/timestampdiff.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/addtime.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/bug3334_ceil.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/bug3506.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/bug3584.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/bug3788.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/CAST.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/concat_ws.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/CONVERT.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/degrees.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/from_days.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/from_unixtime.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/insert.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/LOG2.NS.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/makedate.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/NULLIF.DM.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/RADIANS.NS.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/repeat.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/SUBSTRING_INDEX.sql
Ref failed - working_tpch1/qa_fe_cnxFunctions/subtime.sql
Ref failed - working_tpch1/qa_fe_postProcessedFunctions/CONCAT_WS.SM.sql
Ref failed - working_tpch1/qa_fe_postProcessedFunctions/LOG2.NS.sql
Ref failed - working_tpch1/qa_fe_postProcessedFunctions/SEC_TO_TIME.NS.sql
Local failed - working_tpch1/view/view.sql
Local failed - working_tpch1_calpontonly/misc/bug3513.sql

 Comments   
Comment by Dipti Joshi (Inactive) [ 2016-08-30 ]

Is this passing now hill

Comment by Andrew Hutchings (Inactive) [ 2016-09-26 ]

note: a majority of the ref files are bad. They look like they were generated with a different data set. Comparing with a refDB for now, but they should be regenerated after this.

Comment by Andrew Hutchings (Inactive) [ 2016-09-26 ]

All my notes:

Compare failed - working_tpch1/aggregation/group_concat.sql - Concat of long strings appears to be truncated at 512 characters
 
 
 
Ref failed - working_tpch1/misc/bug2954.sql - First query in test now returns an error in MariaDB and Columnstore. Either remove or update ref
 
Ref failed - working_tpch1/misc/bug3481.sql - Test is columnstore only so fails RefDB
 
Ref failed - working_tpch1/misc/bug4827.sql - Case sensitivity error in test table name
 
Compare failed - working_tpch1/misc/bug2584.sql - The following query needs an order by in the test:
 
select x1.varchar_4, y1.char_8 from table100_char as x1, table10_char as
 
y1 where x1.varchar_4 = y1.char_8;
 
Compare failed - working_tpch1/misc/bug2892.sql - InfiniDB had different week number behaviour to MariaDB. This test will always fail whilst Columnstore retains this behaviour.
 
Compare failed - working_tpch1/misc/bug3475.sql - Precision error in MariaDB (ColumnStore actually handles this better)
 
Compare failed - working_tpch1/misc/bug3881.sql - ColumnStore returns results in different order. I think the queries need fixing
 
 
 
Compare failed - working_tpch1/postprocess/func_3.sql - InfiniDB had different week number behaviour to MariaDB. This test will always fail whilst Columnstore retains this behaviour.
 
Compare failed - working_tpch1/postprocess/order_by_func_bug.sql - InfiniDB had different week number behaviour to MariaDB. This test will always fail whilst Columnstore retains this behaviour.
 
 
 
Ref failed - working_tpch1/qa_fe_cnxFunctions/bug3333.sql - Test is columnstore only so fails RefDB
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/bug3334_ceil.sql - Precision error, I think ColumnStore is generating better results than MariaDB
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/bug3506.sql - Bug: Random seeding behaves differently to MariaDB for negative seed values
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/bug3584.sql - Bug: time(0) doesn't compare to '0' (in quotes). Bug2: compare with '074000' returns wrong result, minutes appear to be truncated.
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/bug3788.sql - Bug: reverse() function puts a '+' before the 'e'
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/CAST.sql - Bug: cast() 0 to date / datetime returns NULL instead of 0 date/datetime
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/concat_ws.sql - MCOL-271
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/CONVERT.sql - Bug convert() 0 to date / datetime returns NULL instead of 0 date/datetime
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/degrees.sql - Precision error
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/EXTRACT.DM.sql - InfiniDB had different week number behaviour to MariaDB. This test will always fail whilst Columnstore retains this behaviour.
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/from_unixtime.sql - Bug: datetime to int returns junk. Also: test needs microsecond support
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/insert.sql - Bug: INSERT() function misbehaves when 'pos' or 'len' parameters are out of range
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/LOG2.NS.sql - Precision error
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/makedate.sql - Bug: year(0) returns NULL
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/MONTHNAME.DS.sql - Bug: MONTHNAME() to int should equal 0, currently equals month number
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/RADIANS.NS.sql - Precision error
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/repeat.sql - MCOL-271
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/SUBSTRING_INDEX.sql - MCOL-271
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/subtime.sql - Bug: subtime off by one second for select cdate, subtime(cdate, '66 12:12:12') from datatypetestm;
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/WEEK.DM.sql - InfiniDB had different week number behaviour to MariaDB. This test will always fail whilst Columnstore retains this behaviour.
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/WEEK.DS.sql - InfiniDB had different week number behaviour to MariaDB. This test will always fail whilst Columnstore retains this behaviour.
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/YEARWEEK.DM.sql - InfiniDB had different week number behaviour to MariaDB. This test will always fail whilst Columnstore retains this behaviour.
 
Compare failed - working_tpch1/qa_fe_cnxFunctions/YEARWEEK.DS.sql - InfiniDB had different week number behaviour to MariaDB. This test will always fail whilst Columnstore retains this behaviour.
 
 
 
Compare failed - working_tpch1/qa_fe_postProcessedFunctions/CONCAT_WS.SM.sql - MCOL-271
 
Compare failed - working_tpch1/qa_fe_postProcessedFunctions/LOG2.NS.sql - Precision error
 
Compare failed - working_tpch1/qa_fe_postProcessedFunctions/SEC_TO_TIME.NS.sql - Requires microsecond support
 
Compare failed - working_tpch1/qa_fe_postProcessedFunctions/WEEK.DM.sql - InfiniDB had different week number behaviour to MariaDB. This test will always fail whilst Columnstore retains this behaviour.
 
Compare failed - working_tpch1/qa_fe_postProcessedFunctions/WEEK.DS.sql - InfiniDB had different week number behaviour to MariaDB. This test will always fail whilst Columnstore retains this behaviour.
 
Compare failed - working_tpch1/qa_fe_postProcessedFunctions/WEEKOFYEAR.DS.sql - InfiniDB had different week number behaviour to MariaDB. This test will always fail whilst Columnstore retains this behaviour.
 
Compare failed - working_tpch1/qa_fe_postProcessedFunctions/YEARWEEK.DM.sql - InfiniDB had different week number behaviour to MariaDB. This test will always fail whilst Columnstore retains this behaviour.
 
Compare failed - working_tpch1/qa_fe_postProcessedFunctions/YEARWEEK.DS.sql - InfiniDB had different week number behaviour to MariaDB. This test will always fail whilst Columnstore retains this behaviour.
 
 
 
Compare failed - working_tpch1/view/view.sql - Bug: bad NULL match when VIEW used in subquery

Bugs to be filed

Comment by Andrew Hutchings (Inactive) [ 2016-09-26 ]

Note that I haven't tried working_tpch1_calpontonly yet as described in the bug description.

Comment by Andrew Hutchings (Inactive) [ 2016-09-26 ]

Update: all of working_tpch1_calpontonly passed.

Comment by Andrew Hutchings (Inactive) [ 2016-10-03 ]

Many of the regression issues have been fixed. The remaining issues are to do with the difference between InfiniDB/ColumnStore and InnoDB/Aria. I have fixed the regression suite for ref file issues.

When enabling all the working_tpch1 in test001 the only failing test now is working_tpch1_compareLogOnly/partitionOptimization/bug5848.sql which is covered by MCOL-339. Since that is being handled separately I'm closing this bug.

Comment by Andrew Hutchings (Inactive) [ 2016-10-03 ]

Chose PrimProc for this as most of the problems were there. See my previous comment for resolution details.

Generated at Thu Feb 08 02:17:57 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.