Sergey Konoplev <gray.ru@gmail.com> writes:
> On Tue, Oct 7, 2014 at 10:16 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> Sergey Konoplev <gray.ru@gmail.com> writes:
>>> BTW, where can I find a list of type1->type2 pairs that doesn't
>>> require full table lock for conversion?
>> There aren't any. Sometimes you can skip a table rewrite, but that
>> doesn't mean that a lesser lock is possible.
> Oh, sorry, it was a typo, I meant "that doesn't require a full table rewrite".
Roughly speaking it's the pairs that have a binary (WITHOUT FUNCTION)
coercion according to pg_cast, although we have special logic for a few
cases such as varchar(M) -> varchar(N).
regards, tom lane