[erlang-questions] mnesia:transform_table

[Igor Ribeiro Sucupira]

If you are still modelling the tables, you might also want to consider
some approach to avoid running mnesia:transform_table/3, since it
needs to rewrite all the records.

For example, as described here:
http://igorrs.blogspot.com/2010/05/mnesia-one-year-later-part-2.html#key_value

And discussed in this thread (from the third post on):
http://forum.trapexit.org/viewtopic.php?p=55510&sid=cc3aec96ba9387b61b5d192ccd1de1a6

Good luck.
Igor.

On Wed, May 26, 2010 at 2:09 AM, Chris Hicks
<silent_vendetta@REDACTED> wrote:
>
> So when updating a record definition the updated record, when reloaded, is not compatible with the old version of the record, as far as I understand it. Updating a Mnesia table to handle the new definition is straightforward enough but, my question revolves around the rest of the running system. Mainly I want to know what is the general (if there is one) process to updating a running system in this manner?
> Should you create a function which locks the whole table, reloads the code and updated all the definitions at the same time? What happens to the rest of the running system when you update a record definition while a process is in the middle of working with that old record?
> _________________________________________________________________
> The New Busy think 9 to 5 is a cute idea. Combine multiple calendars with Hotmail.
> http://www.windowslive.com/campaign/thenewbusy?tile=multicalendar&ocid=PID28326::T:WLMTAGL:ON:WL:en-US:WM_HMP:042010_5



-- 
"The secret of joy in work is contained in one word - excellence. To
know how to do something well is to enjoy it." - Pearl S. Buck.