>>>>> "Ryan" == Ryan Mills <rmills@stripped> writes:
Ryan> Is it possible that a priority for this release could incorporate the
Ryan> development of AUTO_INCREMENT columns in HEAP tables and also the ability to
Ryan> use an index to do an ORDER BY in a HEAP table. Optimisation of the heap
Ryan> type tables would be brilliant, as the powerful combination of a SQL server
Ryan> plus shared hash tables is the primary reason for using MySQL ... It saves
Ryan> having to play around with in memory hash tables in C++, and the faster
Ryan> indexing and better functionality with AUTO_INCREMENTs the better!
Ryan> Many thanks and happy new year,
Ryan> Ryan Mills
Adding AUTO_INCREMENT to HEAP is quite trivial. On the other hand
adding ORDER BY on HEAP is impossible as long as we are using hashing
for key lookups; Hashing is faster to use for lookups but as the keys
are not stored in sorted order one can't use them for ORDER BY.
We will in the future add other key types to HEAP tables, but this is
more likely come in 4.1 or later.
|• Version 4.0||Ryan Mills||4 Jan|
| • Version 4.0||Michael Widenius||4 Jan|