
Originally Posted by
cjard
i've only ever seen rowid increment for the data work that ive been doing and its been quite a reliable thing that for two rows inserted in the same second, the rowid has been able to order them in order of creation.
Then I suppose you've never ever observed how new rows can take place that has been releeased by deleted rows from the same table, thus obtaining a "lower" ROWIDs compared to some rows that were inserted before them? Or new table extents beeng allocated in the tablespace in such location that all rows that will end in that extent in the future will result in ROWIDs that are lower than any currently existing table's row? Or that ROWIDs of the existing rows can change for various reasons, thus making the insertion time even less corelated to the row's ROWID?
In short, as I've allready said: by comparing ROWIDs of two rows, you can definitely not conclude which of those two rows was inserted before the other one. All you can do based on their ROWIDs is *to guess*.
Jurij Modic
ASCII a stupid question, get a stupid ANSI
24 hours in a day .... 24 beer in a case .... coincidence?