Remove hypertable parent from query plans #1007
Conversation
| -> Seq Scan on _hyper_1_1_chunk | ||
| (5 rows) | ||
| QUERY PLAN | ||
| ------------------------------------------------------------------------------------- |
There was a problem hiding this comment.
This plan seems very odd. Why do we have a merge append for only one child
There was a problem hiding this comment.
well that can happen cause postgres doesnt know before how many children it will end up with
Codecov Report
@@ Coverage Diff @@
## master #1007 +/- ##
=========================================
Coverage ? 90.88%
=========================================
Files ? 89
Lines ? 10258
Branches ? 0
=========================================
Hits ? 9323
Misses ? 935
Partials ? 0
Continue to review full report at Codecov.
|
Codecov Report
@@ Coverage Diff @@
## master #1007 +/- ##
=========================================
- Coverage 91.87% 91.78% -0.1%
=========================================
Files 91 91
Lines 10806 10793 -13
=========================================
- Hits 9928 9906 -22
- Misses 878 887 +9
Continue to review full report at Codecov.
|
svenklemm
force-pushed
the
no_parent
branch
2 times, most recently
from
9db0d30
to
c26933e
Compare
test/expected/rowsecurity-11.out
Outdated
| -> Seq Scan on t1 | ||
| Filter: ((a % 2) = 0) | ||
| (3 rows) | ||
| QUERY PLAN |
There was a problem hiding this comment.
we probably do want to add some data to this table so that we can see the filter correctly applied. Right now these test no longer test anything.
test/expected/rowsecurity-11.out
Outdated
| -----------+---------+--------- | ||
| (0 rows) | ||
|
|
||
| ERROR: cursor "current_check_cursor" is not a simply updatable scan of table "current_check" |
There was a problem hiding this comment.
These look bad. Why the change?
svenklemm
force-pushed
the
no_parent
branch
19 times, most recently
from
39e17bb
to
f453c63
Compare
svenklemm
force-pushed
the
no_parent
branch
2 times, most recently
from
92c073b
to
cc87060
Compare
Since the parent table of a hypertable cannot contain any rows including it in the plan is not necessary. PostgreSQL might also not have good statistics for the main table as autoanalyze is triggered by inserts which never happen to the main table, leading to worse plans. This patch adds a new guc enable_constraint_exclusion to disable planner constraint exclusion. Constraint exclusion will not work with cursors and WHERE CURRENT OF so the new guc will allow you to disable that feature. This patch also fixes constraint exclusion for JOINs and sort optimization for queries with chunks_in.
svenklemm
changed the title
Since the parent table of a hypertable cannot contain
any rows including it in the plan is not necessary. PostgreSQL might
also not have good statistics for the main table as autoanalyze is
triggered by inserts which never happen to the main table, leading to
worse plans.
This patch adds a new guc enable_constraint_exclusion to disable
planner constraint exclusion. Constraint exclusion will not work with
cursors and WHERE CURRENT OF so the new guc will allow you to disable
that feature.
This patch also fixes constraint exclusion for JOINs and sort
optimization for queries with chunks_in.