Re: Conflict handling for COPY FROM
| От | Alvaro Herrera |
|---|---|
| Тема | Re: Conflict handling for COPY FROM |
| Дата | |
| Msg-id | 20190714164001.GA21895@alvherre.pgsql обсуждение исходный текст |
| Ответ на | Re: Conflict handling for COPY FROM (Surafel Temesgen <surafel3000@gmail.com>) |
| Ответы |
Re: Conflict handling for COPY FROM
|
| Список | pgsql-hackers |
I think making ERROR a reserved word is a terrible idea, and we don't
need it for this feature anyway. Use a new option in the parenthesized
options list instead.
error_limit being an integer, please don't use it as a boolean:
if (cstate->error_limit)
...
Add an explicit comparison to zero instead, for code readability.
Also, since each error decrements the same variable, it becomes hard to
reason about the state: at the end, are we ending with the exact number
of errors, or did we start with the feature disabled? I suggest that
it'd make sense to have a boolean indicating whether this feature has
been requested, and the integer is just the remaining allowed problems.
Line 3255 or thereabouts contains an excess " char
The "warn about it" comment is obsolete, isn't it? There's no warning
there.
--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
В списке pgsql-hackers по дате отправления: