Hi,<br /> The Current architecture of Updates in PostgreSQL is <br />1) Make a select query out of update. It
involvesa READ lock/BUFFER_SHARE<br />2) Get the tupleid<br />3) Goto the buffer containing the tupleid, make a
BUFFER_EXCLUSIVElock on it <br />4) update it<br />5) Repeat the above process for subsequent rows<br /><br />I propose
tochange this row-by-row approach, when it is a full table update. I plan to send a extra flag(which will be set for
Fulltable Deletes/Updates). this would make the access method directly acquire the exclusive lock and update the
existingrecord. <br /><br />For Deletes this is simple. But for updates, the projection tuple has to be made before
re-insertingit. So there will be a list of Heap tuples stored in memory for each page getting updated. these tuples
willbe inserted after the deletion part of update is done. This is just a rough design. I may get involved in a detail
designonce i get a nod from the mailing list community. <br /><br />Thanks,<br />Gokul.<br />
