Details
-
Bug
-
Status: Open (View Workflow)
-
Minor
-
Resolution: Unresolved
-
10.5, 10.6, 10.11, 11.0(EOL), 11.1(EOL), 11.2(EOL), 11.4
-
None
Description
When a new batch of values is read from a sequence, TABLE_STATISTICS counts the sequence row as changed twice. It wasn't so initially, but started happening after this commit in 10.5.2:
commit 91ab42a823b244a3d4b051ab79701b7a552f274a
|
Author: Monty
|
Date: Tue Jan 28 23:23:51 2020 +0200
|
|
Clean up and speed up interfaces for binary row logging
|
set @userstat.save= @@userstat; |
set global userstat= 1; |
|
create sequence s; |
flush table_statistics;
|
select nextval(s); |
query_vertical show table_statistics;
|
|
# Cleanup
|
drop table s; |
set global userstat= @userstat.save; |
10.5 8bc32410160265dcd27d595ed610843d08d32034 |
flush table_statistics;
|
select nextval(s);
|
nextval(s)
|
1
|
show table_statistics;
|
Table_schema test
|
Table_name s
|
Rows_read 0
|
Rows_changed 2
|
Rows_changed_x_#indexes 2
|
In MDEV-33151 where the changed rows are detalized as inserts/updates/deletes, it can be seen that the row is counted as inserted once and updated once.