Hi, I'm implementing a shopping system.
---- webware version CVS-20041011 ----
1. I have to migrate a huge amount of product data from several sources
especially about 2000000 products in csv-format(books), a firebird
database(bookshop stock db)into only one stable db.
I want to use postgre for that purpose.
First I wrote a prototype without middlekit and used the ISBN / EAN
as my primary key for the products table. After writing my Classes.csv I
realized, that middlekit automatically creates a sequence and a primary key
in PostgreSQLGenerator.py:
def primaryKeySQLDef(self, generator):
return "\t%s integer not null primary key default
nextval('%s'),\n" % (self.sqlSerialColumnName(), self.seqName())
Question:
Would it be be more useful to use the implemented primary key format
and hold the ean/isbn as a "long type secondary key" / create an index
on that collumn or
How can I tell midllekit to use a long key, not creating
2. It would be nice having more control about the table indexing
procedure. I'm thinking about write some new code, but in the moment I
don't know where to begin, how middlekit could work with these
modifications, and if someone of you have done some work about it...
thanks
Stefan
|