[Grok-dev] Re: 0.14 wishlist items

Wichert Akkerman wichert at wiggy.net
Fri Jun 13 12:10:00 EDT 2008


Previously Sebastian Ware wrote:
> The use case the way I see it is to allow the same data export and  
> import that can be done with an RDB, nothing more. Granted the  
> hierarcical structure with containers makes it slightly more tricky.  
> Anyway, I still have my "add value index type with tests" assignment  
> to finish... :p still, this is a feature I need.

Keep in mind that 'data export' as I think you intend it is something
very different than 'zodb export' other people refer to:

* ZODB dumps (like the .zexp dumps Zope2 produces) are a full
  representation of all data in the system. This means it is highly
  tied to the specific versions of packages you have installed and
  lots of implementation details. Such a dump is useless for most
  purposes except backup and restore or moving data between two
  systems that are completely equivalent.

* data dumps that contain more 'normal data'. With that I mean things
  like objects and their data as defined in a schema. This is data
  that has a well defined structure and can easily be handled by
  standard tools.

Tools such as collective.transmogrify try to handle the second part by
giving you tools to flexibly move such data around and transform it in
any way you want. 

It is important that you keep in mind which type of data export/import
you want handle before making a decision and painting yourself in a
corner.

Wichert.

-- 
Wichert Akkerman <wichert at wiggy.net>    It is simple to make things.
http://www.wiggy.net/                   It is hard to make things simple.


More information about the Grok-dev mailing list