Re: INSERT ... ON CONFLICT IGNORE (and UPDATE) 3.0

On 2015-04-21 16:57:45 +0200, Andres Freund wrote:
> * I still think it's unacceptable to redefine
>   XLOG_HEAP_LAST_MULTI_INSERT as XLOG_HEAP_SPECULATIVE_TUPLE like you
>   did. I'll try to find something better.

I think we should "just" split this into different flag values for
insert/update/delete.

I.e. something like

/* flags for heap insert and multi insert */
#define XLH_INSERT_ALL_VISIBLE_CLEARED
#define XLH_INSERT_LAST_MULTI_INSERT
#define XLH_INSERT_IS_SPECULATIVE
#define XLH_INSERT_CONTAINS_NEW_TUPLE

/* flags for update */
#define XLH_UPDATE_OLD_ALL_VISIBLE_CLEARED
#define XLH_UPDATE_NEW_ALL_VISIBLE_CLEARED
#define XLH_UPDATE_CONTAINS_OLD_TUPLE
#define XLH_UPDATE_CONTAINS_OLD_KEY
#define XLH_UPDATE_CONTAINS_NEW_TUPLE
#define XLH_UPDATE_PREFIX_FROM_OLD
#define XLH_UPDATE_SUFFIX_FROM_OLD

/* flags for delete */
#define XLH_DELETE_ALL_VISIBLE_CLEARED
#define XLH_DELETE_CONTAINS_OLD_TUPLE
#define XLH_DELETE_CONTAINS_OLD_KEY

Greetings,

Andres Freund