F.50. sr_plan


sr_plan is incompatible with extensions that use post_parse_analyze_hook, such as pg_stat_statements. Do not use sr_plan together with such extensions.

F.50.1. Rationale

sr_plan is an extension which allows to save query execution plans and use these plans for all repetitions of same query, instead of optimizing identical query again and again.

sr_plan looks like Oracle Outline system. It can be used to lock the execution plan. It is necessary if you do not trust the planner or able to form a better plan.

Typically, DBA would play with queries interactively, and save their plans and then enable use of saved plans for the queries, where predictable response time is essential.

Then application which uses these queries would use saved plans.

F.50.2. Installation

In your db:


and modify your postgresql.conf:

shared_preload_libraries = 'sr_plan'

It is essential that library is preloaded during server startup, because use of saved plans is enabled on per-database basis and doesn't require any per-connection actions.

F.50.3. Usage

If you want to save the query plan is necessary to set the variable:

set sr_plan.write_mode = true;

Now plans for all subsequent queries will be stored in the table sr_plans, until this variable is set to false. Don't forget that all queries will be stored including duplicates. Making an example query:

select query_hash from sr_plans where query_hash=10;

Disable saving the query:

set sr_plan.write_mode = false;

Now verify that your query is saved:

select query_hash, enable, valid, query, explain_jsonb_plan(plan) from sr_plans;

 query_hash | enable | valid |                        query                         |                 explain_jsonb_plan                 
 1783086253 | f      | t     | select query_hash from sr_plans where query_hash=10; | Bitmap Heap Scan on sr_plans                      +
            |        |       |                                                      |   Recheck Cond: (query_hash = 10)                 +
            |        |       |                                                      |   ->  Bitmap Index Scan on sr_plans_query_hash_idx+
            |        |       |                                                      |         Index Cond: (query_hash = 10)             +
            |        |       |                                                      | 

Note use of explain_jsonb_plan function, that allows you to visualize execution plan in the similar way as EXPLAIN command does.

In the database plans are stored as jsonb. By default, all the newly saved plans are disabled, you need enable it manually:

To enable use of the saved plan

update sr_plans set enable=true where query_hash=1783086253;

(1783086253 for example only) After that, the plan for the query will be taken from the sr_plans table.

In addition sr plan allows you to save a parameterized query plan. In this case, we have some constants in the query that, as we know, do not affect plan.

During plan saving mode we can mark these constants as query parameters using a special function _p (anyelement). For example:

=>create table test_table (a numeric, b text);
=>insert into test_table values (1,'1'),(2,'2'),(3,'3');
=> set sr_plan.write_mode = true;
=> select a,b  from test_table where a = _p(1);
 a | b
 1 | 1
(1 row)

=> set sr_plan.write_mode = false;

Now plan for query from our table is saved with parameter. So, if we enable saved plan in this table, this plan would be used for query with any value for a, as long as this value is wrapped with _p() function.

=>update sr_plans set enable = true where query=
  'select a,b from test_table where a = _p(1)';
-- These queries would use saved plan

=>select a,b from test_table where a = _p(2);
 a | b
 2 | 2
(1 row)

=>select a,b from test_table where a = _p(3);
 a | b
 3 | 3
(1 row)

-- This query wouldn't use saved plan, because constant is not wrapped
-- with _p()

=>select a,b from test_table where a = 1;
 a | b
 1 | 1
(1 row)