Re: json(b)_array_elements use causes very large memory usage when also referencing entire json document

On Fri, Oct 6, 2017 at 11:48 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:

> Lucas Fairchild-Madar <lucas(dot)madar(at)gmail(dot)com> writes:
> > I'm experiencing this on 9.6.
>
> Oh, hmm, it's something else in that case. Can you put together a
> more self-contained test case? This sort of thing isn't usually
> very data-dependent, so you could probably make a test case with
> dummy data.
>
> regards, tom lane
>

Sure, here's a bash script. Assumes you have PGDATABASE, etc set.
Verified this also blows up on pg 10. I haven't verified the new patch.

#!/bin/bash

ITEMS=100000
FILE=kaboom.sql

echo -n "{\"id\":\"foo\",\"items\":[" > $FILE
for i in $(seq 1 $ITEMS); do
echo -n "{\"item\":$i,\"data\":\"The quick brown cow jumps over the lazy
sloth\"}," >> $FILE
done
echo "{\"dummy\":true}]}" >> $FILE

psql -e -c "drop table if exists kaboom; create table kaboom(data jsonb);"
psql -e -c "\\copy kaboom from kaboom.sql"
echo "this works"
psql -qAtc "select jsonb_array_elements(data->'items') from kaboom;" | wc
echo "prepare to swap"
psql -qAtc "select data->'id', jsonb_array_elements(data->'items') from
kaboom;" | wc