mercredi 13 août 2014

php - Django créant des tables différentes pour le même modèle - Stack Overflow


Earilier i was using PHP without any framework, for a service oriented web application, where i had to store a large volume of data to be stored, which could be easily removed and simple for making a backup; so i wrote code to store the same data in different tables of same architecture for every month, so that can easily run drop table query or create backup of table rather than running query on the whole table.


+------------------+---------------------------+
| data_2014_01_01 | will store data of jan |
| data_2014_02_01 | will store data of feb |
| data_2014_03_01 | will store data of march |
+------------------+---------------------------+

and so on.


Since i am migrating the code to django, how can i get the same thing done, because django ORM makes a single table, for every data item to be stored here, hence deleting the useless older record which can now only be identified by timestamp is taking too much time and taking its backup is also painfull, because the size of table is very big.


Please advice me some techniques here, any ideas will be welcomed.


Thanks in advance.




Here's an option.


Since you are moving to Django, I'd suggest to follow the best practices and start using South right away:



South is a tool to provide consistent, easy-to-use and database-agnostic migrations for Django applications.



Also, I'd combine all of your tables into single one with a timestamp field - this way it would be easier for the application to "talk" to your data.


Then, besides setting up South, there are steps that you need to do only once:



  • create your schema - define your model, create an initial schema migration and run it.

  • write a data migration that will move the data from multiple data_%Y_%m_%d tables into the newly created by your schema migration table. Run your data migration.




Also take a look at django-parting project - it doesn't look very mature and ready to use, but, at least, see how they managed multiple tables per model (basically it's a bit of Model Manager "magic").


Hope that helps.



Earilier i was using PHP without any framework, for a service oriented web application, where i had to store a large volume of data to be stored, which could be easily removed and simple for making a backup; so i wrote code to store the same data in different tables of same architecture for every month, so that can easily run drop table query or create backup of table rather than running query on the whole table.


+------------------+---------------------------+
| data_2014_01_01 | will store data of jan |
| data_2014_02_01 | will store data of feb |
| data_2014_03_01 | will store data of march |
+------------------+---------------------------+

and so on.


Since i am migrating the code to django, how can i get the same thing done, because django ORM makes a single table, for every data item to be stored here, hence deleting the useless older record which can now only be identified by timestamp is taking too much time and taking its backup is also painfull, because the size of table is very big.


Please advice me some techniques here, any ideas will be welcomed.


Thanks in advance.



Here's an option.


Since you are moving to Django, I'd suggest to follow the best practices and start using South right away:



South is a tool to provide consistent, easy-to-use and database-agnostic migrations for Django applications.



Also, I'd combine all of your tables into single one with a timestamp field - this way it would be easier for the application to "talk" to your data.


Then, besides setting up South, there are steps that you need to do only once:



  • create your schema - define your model, create an initial schema migration and run it.

  • write a data migration that will move the data from multiple data_%Y_%m_%d tables into the newly created by your schema migration table. Run your data migration.




Also take a look at django-parting project - it doesn't look very mature and ready to use, but, at least, see how they managed multiple tables per model (basically it's a bit of Model Manager "magic").


Hope that helps.


0 commentaires:

Enregistrer un commentaire