How to get "memory usage" on postgres CTE query plan? - postgresql

Question

I'm comparing different queries options in order to improve perfomance and I cannot understand why my CTE query plan does not show the "memory usage" option when I run it with explain analyze.
Here's my query:
EXPLAIN ANALYZE WITH
CTE AS (
SELECT
id,
jsonb_agg(CASE WHEN each_result->>'another_key' = 'A'
THEN jsonb_set(each_result, '{another_key}', '"B"', false)
ELSE each_result
END
) AS result_updated
FROM my_table, jsonb_array_elements(column->'key') as result
GROUP BY id
)
UPDATE my_table
SET column = jsonb_set(column, '{each_result}', (SELECT result_updated FROM CTE WHERE CTE.id = my_table.id), false)
;
And here's the query plan result:
QUERY PLAN
-------------------------------------------------------------------------------------------------------
Update on custom_sequence (cost=3305.32..25976.06 rows=1003 width=83) (actual time=4071.870..4071.870 rows=0 loops=1)
CTE cte
-> HashAggregate (cost=3292.78..3305.32 rows=1003 width=48) (actual time=536.656..618.699 rows=1003 loops=1)
Group Key: my_table_1.id
-> Nested Loop (cost=0.01..2039.04 rows=100300 width=48) (actual time=0.323..59.200 rows=78234 loops=1)
-> Seq Scan on my_table my_table_1 (cost=0.00..33.03 rows=1003 width=34) (actual time=0.023..0.289 rows=1003 loops=1)
-> Function Scan on jsonb_array_elements step (cost=0.01..1.00 rows=100 width=32) (actual time=0.044..0.049 rows=78 loops=1003)
-> Seq Scan on my_table (cost=0.00..22670.74 rows=1003 width=83) (actual time=629.743..3701.285 rows=1003 loops=1)
SubPlan 2
-> CTE Scan on cte (cost=0.00..22.57 rows=5 width=32) (actual time=1.992..3.482 rows=1 loops=1003)
Filter: (id = my_table.id)
Rows Removed by Filter: 1002
Planning time: 0.458 ms
Execution time: 4081.778 ms
(14 rows)
My another option is an UPDATE FROM:
EXPLAIN ANALYZE UPDATE my_table
SET column = jsonb_set(column, '{each_result}', q2.result_updated, false)
FROM (
SELECT
id,
jsonb_agg(CASE WHEN each_result->>'anoter_key' = 'A'
THEN jsonb_set(each_result, '{another_key}', '"B"', false)
ELSE each_result
END
) AS result_updated
FROM my_table, jsonb_array_elements(column->'key') AS result
GROUP BY id
) q2
WHERE q2.id = my_table.id
;
When I run query plan for this query, I got the "memory usage" info :
QUERY PLAN
----------------------------------------------------------------------------------------------------------------------------
Update on my_table (cost=3748.30..4223.44 rows=1001 width=153) (actual time=992.003..992.003 rows=0 loops=1)
-> Hash Join (cost=3748.30..4223.44 rows=1001 width=153) (actual time=633.204..758.817 rows=1004 loops=1)
Hash Cond: (my_table.id = q2.id)
-> Seq Scan on my_table (cost=0.00..460.01 rows=1001 width=67) (actual time=0.031..0.550 rows=1004 loops=1)
-> Hash (cost=3735.79..3735.79 rows=1001 width=120) (actual time=632.951..632.951 rows=1004 loops=1)
Buckets: 1024 (originally 1024) Batches: 2 (originally 1) Memory Usage: 32791kB
-> Subquery Scan on q2 (cost=3713.26..3735.79 rows=1001 width=120) (actual time=537.310..605.349 rows=1004 loops=1)
-> HashAggregate (cost=3713.26..3725.78 rows=1001 width=48) (actual time=537.306..604.269 rows=1004 loops=1)
Group Key: my_table_1.id
-> Nested Loop (cost=0.01..2462.01 rows=100100 width=48) (actual time=0.373..47.725 rows=78312 loops=1)
-> Seq Scan on my_table my_table_1 (cost=0.00..460.01 rows=1001 width=34) (actual time=0.014..0.483 rows=1004 loops=1)
-> Function Scan on jsonb_array_elements step (cost=0.01..1.00 rows=100 width=32) (actual time=0.036..0.039 rows=78 loops=1004)
Planning time: 0.924 ms
Execution time: 998.818 ms
(14 rows)
How can I get the memory usage information on my CTE query too?