Re: jsonpath
От | Tomas Vondra |
---|---|
Тема | Re: jsonpath |
Дата | |
Msg-id | eebc218c-7aaa-c425-6ba0-78b28c751830@2ndquadrant.com обсуждение исходный текст |
Ответ на | Re: jsonpath (Alexander Korotkov <a.korotkov@postgrespro.ru>) |
Ответы |
Re: jsonpath
|
Список | pgsql-hackers |
On 1/28/19 1:22 AM, Alexander Korotkov wrote: > On Sun, Jan 27, 2019 at 1:50 PM Alexander Korotkov > <a.korotkov@postgrespro.ru> wrote: >> On Sat, Jan 26, 2019 at 4:27 AM Alexander Korotkov >> <a.korotkov@postgrespro.ru> wrote: >>> >>> On Wed, Jan 23, 2019 at 8:01 AM Alexander Korotkov >>> <a.korotkov@postgrespro.ru> wrote: >>>> Finally, I'm going to commit this if no objections. >>> >>> BTW, I decided to postpone commit for few days. Nikita and me are >>> still working on better error messages. >> >> Updated patchset is attached. This patchset includes: >> >> * Improved error handling by Nikita, revised by me, >> * Code beautification. >> >> So, I'm going to commit this again. This time seriously :) > > I'm really sorry for confusing people, but I've one more revision. > This is my first time attempting to commit such a large patch. > I'm not sure what you're apologizing for, it simply shows how careful you are when polishing such a large patch. > Major changes are following: > * We find it ridiculous to save ErrorData for possible latter throw. > Now, we either throw an error immediately or return jperError. That > also allows to get rid of unwanted changes in elog.c/elog.h. OK. > * I decided to change behavior of jsonb_path_match() to throw as less > errors as possible. The reason is that it's used to implement > potentially (patch is pending) indexable operator. Index scan is not > always capable to throw as many errors and sequential scan. So, it's > better to not introduce extra possible index scan and sequential scan > results divergence. > Hmmm, can you elaborate a bit more? Which errors were thrown before and are not thrown with the current patch version? regards -- Tomas Vondra http://www.2ndQuadrant.com PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
В списке pgsql-hackers по дате отправления: