Re: [HACKERS] UPDATE of partition key

Поиск
Список
Период
Сортировка
От amul sul
Тема Re: [HACKERS] UPDATE of partition key
Дата
Msg-id CAAJ_b97B86ouFnNV9=uTYi6yEYx4hM53YYAriozzEWTAKAOOeQ@mail.gmail.com
обсуждение исходный текст
Ответ на Re: [HACKERS] UPDATE of partition key  (Amit Khandekar <amitdkhan.pg@gmail.com>)
Ответы Re: [HACKERS] UPDATE of partition key
Список pgsql-hackers
Дерево обсуждения
On Wed, Sep 20, 2017 at 9:27 PM, Amit Khandekar <amitdkhan.pg@gmail.com> wrote:
On 20 September 2017 at 00:06, Robert Haas <robertmhaas@gmail.com> wrote:
> On Fri, Sep 15, 2017 at 7:25 AM, Amit Khandekar <amitdkhan.pg@gmail.com> wrote:
>> [ new patch ]

  86 -           (event == TRIGGER_EVENT_UPDATE && !trigdesc->trig_update_after_row))
  87 +           (event == TRIGGER_EVENT_UPDATE && !trigdesc->trig_update_after_row) ||
  88 +           (event == TRIGGER_EVENT_UPDATE && (oldtup == NULL || newtup == NULL)))
  89             return;
  90     }


Either of oldtup or newtup will be valid at a time & vice versa.  Can we improve
this check accordingly?

For e.g.: 
(event == TRIGGER_EVENT_UPDATE && )(HeapTupleIsValid(oldtup) ^ ItemPointerIsValid(newtup)))))


 247 
 248 +   /*
 249 +    * EDB: In case this is part of update tuple routing, put this row into the
 250 +    * transition NEW TABLE if we are capturing transition tables. We need to
 251 +    * do this separately for DELETE and INSERT because they happen on
 252 +    * different tables.
 253 +    */
 254 +   if (mtstate->operation == CMD_UPDATE && mtstate->mt_transition_capture)
 255 +       ExecARUpdateTriggers(estate, resultRelInfo, NULL,
 256 +                    NULL,
 257 +                    tuple,
 258 +                    NULL,
 259 +                    mtstate->mt_transition_capture);
 260 +
 261     list_free(recheckIndexes);

 267 
 268 +   /*
 269 +    * EDB: In case this is part of update tuple routing, put this row into the
 270 +    * transition OLD TABLE if we are capturing transition tables. We need to
 271 +    * do this separately for DELETE and INSERT because they happen on
 272 +    * different tables.
 273 +    */
 274 +   if (mtstate->operation == CMD_UPDATE && mtstate->mt_transition_capture)
 275 +       ExecARUpdateTriggers(estate, resultRelInfo, tupleid,
 276 +                    oldtuple,
 277 +                    NULL,
 278 +                    NULL,
 279 +                    mtstate->mt_transition_capture);
 280 +

Initially, I wondered that why can't we have above code right after
​ExecInsert()​ & ​ExecIDelete()​ in ​ExecUpdate​ respectively?

We can do that for ExecIDelete() but not easily in the ExecInsert() case,
because ExecInsert() internally searches the correct partition's resultRelInfo
for an insert and before returning to ExecUpdate resultRelInfo is restored
to the old one.  That's why current logic seems to be reasonable for now.
Is there anything that we can do?

Regards,
Amul


В списке pgsql-hackers по дате отправления: