Datamapper DataMapper

Comments and changes to this ticket

• 

  Dan Kubb (dkubb) January 16th, 2009 @ 06:39 PM

  I'm curious to know which storage engines you would like to use, and if you'd considered using migrations to alter them to whatever your preference is?

• You flagged this item as spam.
  

  Rimas Silkaitis January 29th, 2009 @ 01:13 PM

  Engines that I had in mind were percona xtradb (https://launchpad.net/percona-xt... and Infobright (http://www.infobright.org/Downlo.... My needs are along the lines of data warehousing.
  

  More specifically, one of the limitations of Infobright's brighthouse storage engine is that you can't switch from other tables (http://dev.mysql.com/tech-resour... since the ALTER TABLE command doesn't work.

  If we could specify the engine right off the bat outside of migrations that would save the need of having to load dm-migrations when the only need would be to change the storage engine. I'm really not sure how much overhead loading dm-migrations would have, but just a thought.

• 

  Dan Kubb (dkubb) February 21st, 2009 @ 12:47 AM

  • State changed from “unconfirmed” to “not-applicable”

  Unfortunately dm-core isn't going to add an API to adjust the options for a single adapter. I would suggest creating a plugin that allows you to do this, for both auto-migrations and classic migrations.

  We have about 20-30 adapters now, and each backend has dozens if not hundreds of options that can be configured. Most of those options are very specific and can't be abstracted into something reusable. If we add them to the dm-core API we'll quickly end up with something really messy, so I'd prefer that specifics like this be moved into their own plugins.

• You flagged this item as spam.
  

  Rimas Silkaitis February 21st, 2009 @ 09:36 AM

  Sounds good. I'll try writing a plugin for this then. Thanks for looking into my request!

• 

  Dan Kubb (dkubb) February 22nd, 2009 @ 12:58 AM

  Great, thanks for understanding. My primary concern is just keeping the API slim and simple. I truly do wish that storage systems were actually more similar than they are, so we could abstract more of the common parts out, but the more we work with varying systems the less similar they seem.