Subquery magic in PostgreSQL

I have a query:

update product_product 
set (write_date, default_code) = (LOCALTIMESTAMP, 'update') 
where product_tmpl_id in (
    select distinct product_tmpl_id 
    from product_template 
    where type='import');

And it takes 7 hours to complete. However, when I execute the subquery:

select distinct product_tmpl_id 
from product_template 
where type='import';

I get the error:

column “product_tmpl_id” does not exist

There isn’t a column product_tmpl_id in the table product_template.

It was my mistake. The first query should have been:

update product_product set (write_date,default_code) = (LOCALTIMESTAMP,'update') 
where product_tmpl_id in
 (select distinct id from product_template where type='import');

And it onle takes several seconds to run.

So my questions are the following:

• why didn`t the first query fail?
• why took it so long to run?
• what exactly did it do?

The result of select version(); is

PostgreSQL 9.4.3 on x86_64-unknown-linux-gnu, 
compiled by gcc (Debian 4.9.2-10) 4.9.2, 64-bit