Well, we're not going to wait until Genezzo supports all the functions and operators from the SQL standard, because that might take a while. The major missing SQL language features are GROUP BY, ORDER BY, aggregation, and subqueries, and I'd classify them as "nice to have", but not "critical" for beta. Also, capabilities like locking and transactions won't be part of the
base product -- we should be able to package them as external modules, similar to the cluster support in Genezzo::Contrib::Clustered. However, beta users should be able to upgrade to locking and transaction support without "undue pain", where "undue pain" is defined as "full export and import of all data", so we want to make sure that the data dictionary, Havok/SysHook external module support, and disk block and row format are sufficiently robust and extensible to support this type of upgrade. The other outstanding issue is space management, which is too simplistic at present. While Genezzo is fine for loads that are dominated by inserts and updates, it uses space inefficiently if you use lots of deletes (eg you tend to delete and re-insert > 50% of your table). Currently, the only recourse is to drop and recreate the table. The planned fixes for this deficiency also requires some dictionary and block format changes, probably sufficient to require an export and import, so I'd like to get these changes in before going beta.
Jeff
(1)
]
