Modifying TOAST policy will not affect the way existing data is stored ?

Em qua., 27 de out. de 2021 às 16:48, Marcos Pegoraro <marcos@f10.com.br>
escreveu:

Postgres weekly came today with this interesting article:
https://hakibenita.com/sql-medium-text-performance

The problem it explains is when you have no one records being stored on
toast, even if they are toastable. That is because the size of that column
fits on toast_tuple_target size, so it does not toast.
What it talks fits perfectly on some tables of mine, so I want them this
way, but how ?

First I did, as article mentioned ...
alter table ... set (toast_tuple_target=128);
alter table ... alter mycol set storage external;

Then I tried vacuum full, cluster, create another column and update its
value with old column and last test was creating a new table with those
definitions and insert from select old table and none of these methods I
could move data from that column to toast. why ?

pg_column_size gives me what number exactly ? Its size is before
compression or later ?

That table has 2.5 million records, average of pg_column_size is 100
bytes but 100 thousand records have more than 500, 50 thousands have more
than 1.000 using pg_column_size

How can I change existing records from table to toast ? Or I cannot ?

then, more some tests, but no one explain.

table_size - toast_size - toast_percent - table
570.146.816 - 2.105.344 - 0.00369% - original_table
564.060.160 - 2.072.576 - 0.00367% - insert_from_select_from_original
551.553.408 - 21.241.856 - 0.03851% - export_to_script_and_import

So, my original table has only 0,003% of size on its toast table.
If I cluster, vacuum full or create an additional field moving old data,
nothing happens
If I create another table and move data to it with insert from select,
nothing happens too.
Then, if I create an external script and then run that insert of 2.5
million rows, cool, then 10 times more of info were moved to toast. why ?

all tables I did a vacuum full as a last step.