From: <Mar...@si...> - 2001-12-24 11:00:08
|
Dear Chalain, Thank you for your kind response. > 3. Any chance of making the loading of a new database (rather than an > update) faster. I am assuming the current approach is to create the > database and then add the records in over the link. This is rather slow for > large databases, when "all you need do" :) is grab the .pdb and sort it out > later. It also seems to fall over for read only databases. >?? my english is too bad to understand. Sorry. I'll try again. I am trying to transfer a large (>3000 record) database from my main computer to my palm. It takes forever. The reason is that each record is being merged into the database rather than db3 spotting that this is a new database (rather than a synchronisation). If the database is new, there is no need to update pointers every record, it can be done at the end. In addition, if a database is loaded using synchronisation, each record the program has to decide where it goes in the database, but since records are added from front to back each record is added at the end. Searching from the start through the whole database takes a *long time* especially when there are lots of records. This is what I *think* is going on, but I may be wrong. Thanks for your help, Martin |