From: Mark D. <mj...@ge...> - 2009-05-28 21:10:39
|
Rutger, Val and I just had a meeting in which we decided on a plan for how to do this. Either Rutger or I will write a program that dumps out the database tables as a series of SQL 'insert' commands. We will be responsible for making sure that the data is properly escaped. The files will be UTF-8 encoded, so that the Unicode data that is currently in the database will be properly represented. The same program can also emit some SQL commands that set the sequence numbers, grants, or whatever. SDSC has assured us that they can provide enough disk space at SDSC to do this. Files can be copied from SDSC to NESCent with any of several methods, including rsync. Once at NESCent, someone (who?) will be responsible for executing these large SQL batch files to import the data into the Pg database. If done properly, the Unicode data will be transferred faithfully. If this sounds like a bad idea, or if there are unanswered questions, now is the time to speak up. |