Re: Informix and large databases

>We're dissatisfied with our current DBMS and are looking for information >about how Informix would handle the following scenario: > >We have tables that have approx 77 million rows that occasionally get >full batch updates. We have no particular requirement to have our >update transactions logged as we can re-run a job if it fails to finish. > >Our question is two-fold: > >1. Does Informix support some kind of "batch" updating that would allow >us to update single fields from a staging table without dividing the >transaction into parts, and without carrying the overhead of logging? >The update would look something like > >Update master_table >Set master_table.my_field = trans_table.update_field >where master_table.key = trans_table.key > >2. Our server has 6 cpu's. Does Informix take advantage of multiple >cpu's when processing queries, or during initial data loads? Yes - Informix has a batch loader, High Performance Loader, which will do parallel loads. This will also allow the load to be done in an un-logged mode and will then allow the fragments loaded to be attached to the existing table fragments. Also normal queries can easilyl be parallelized at the various levels of the query such as the scan, sort, group, etc... > >We function much more as a data warehouse than as an OLTP system. There >are no "real-time" updates in our system, though we support multiple >concurrent queries. > > >Thanks in advance for any insights. > >RLL Madison Pruet