I have two different applications (two different solutions) using the same database. Each application has its own set of models/tables used in it.
The .NET solution for Application 1 was created a year back and has evolved till today. As part of this, some new models were created, migrations were added and updated in the database (SampleDatabase). This has its own DBContext class.
The .NET solution for Application 2 is a new one created a month back. This application also uses the same database (SampleDatabase). A few new models specific to this application were created, migrations were added and updated in the database. This has its own DBContext class.
Requirement:
Now, there is a requirement where one of the tables created and being used by Application 1 needs to be modified to integrate with Application 2 (i.e., a PK from an Application 2 table has to be placed as a FK in an Application 1 table.).
Challenge
To create a migration for altering the model from Application 1, a reference is needed to the PK table available in the other solution, which is not possible.
To create a migration for altering the existing model from Application 2, the existing model has to be copied to the Application 2 codebase. By doing so, if a migration is added, it generated only a CreateTable statement and not an alter as the DBContext does not know that the table already exists.
Question:
How do we manage to have migrations that is aware of things happening in the database from both the solutions?
Thoughts:
One way that I had in mind is to have a shared library project that can be responsible for all migrations to this database. Requirements for any application can be taken care by this project. But this has to refer the DBContext from both the application which is again a challenge.
Please let me know your thoughts and if anyone has faced such situation in your projects.
Thank You.
Related
Need help! I am new to Core 2.0 and VS 2017 and haven't had much experience with MVC. I have an existing database that I need to use with a core 2.0 project at work. I was able to use the Scaffold-DbContect to initially create the domain models from the existing database.
However, the database developers are making changes to the database and adding new tables. I need to keep my domain models in sync with the database changes that are being made.
The only thing I can find on the internet is how to make changes to the model and update the database schema. However, I need to update the model from changes made to the database.
EF Core works on Code First approach. And You guys are following DB First approach together. So you should make changes in your code and then generate migrations accordingly, Otherwise, it will lead you in trouble.
You can use EF Core Power tool for generating the db changes at code side. But in this case you have to take care while generating migrations from code side.
we have a project running Entity Framework 6.1 and we started using explicit migrations a while ago (in the past we would use automatic migrations) and we have run into the following situation.
I create an explicit migration to create some indexes on fields. I do this in a seperate branch.
A colleague of mine also starts an explicit migration to do some other work in his own branch.
Each branch is code reviewed and merged into the master branch when approved.
But now we noticed that my explicit migration to create the indexes, was created on a different version of the model. Since it's a project with multiple developers, the model is always changing. So if we check what SQL code would be generated to update the database, we see that new columns/tables/... that have been added in the meantime while I was working on my branch are being dropped, that my indexes are then created and afterwards those columns would be added again.
How can we avoid this? What are we doing wrong in our workflow?
With EF 6 each migration has metadata about last state of DB.
In EFCore this is much better done with separate file that has snapshot of DB.
Here are some good practices for Migrations in Team Environments:
https://msdn.microsoft.com/en-us/library/dn481501(v=vs.113).aspx
How to manage Migrations in a project with multiple branches?
Now your situation is pretty specific and I am not sure that any of these procedures has automatic solution for it.
One way I can think of is to have DB model not locally but on a server and that each developer targets that when creating migration.
However in the previous blogs shared DB was not considered best practice.
You would need to figure out some hybrid procedure to comply with every advice.
Good luck...
Is there a more elegant way to maintain a view in Model-First Entity Framework? There's plenty of discussion regarding how to do this using Migrations in Code-First, but of course Migrations are not available when using Model-First.
My current approach:
Run "Generate Database from Model..."
Run the created script
Delete the table EF generates for the view
Create the view manually
The last two steps are done in a post-script, of course.
I'm not expecting an integrated view-designer (although I think that's what MS was originally intending for future releases, before they fell in love with Code-First). For example, if it's possible simply to tell EF not to create that table, it would be a small improvement.
I have an application that uses Entity Framework code first migrations where the application is deployed on two servers both using the same database. Now I have a simple database update where a table and the EF model has a new column/property. I have created the migration and it works fine in a one server scenario.
But how do I deploy this to two servers without downtime? Without EF I would just start out and add the column to the table and then update the servers one by one. The old app would work just fine against the updated database as long as it is a simple change like this. What is the best way to do this in EF? Can I avoid problems in the second, not updated server, while I am updating the first one and the database?
This sounds like a perfect candidate for a Mirrored database, assuming you are using SQL Server.
You'd just apply your migrations to the Principal database and it will take care of the rest behind the scenes.
I'm using Entity Framework 5 model first. Say I've deployed the application and I'd like to upgrade an EF entity with new columns, basically adding columns to the table.
What is the best way to upgrade the existing database without losing data? For example I have a User table that I add two new columns to. If I try to script a schema change the tables will need to be dropped in order to add the new columns. Is there a way to update the tables without needing to recreate them? Thanks!
This may be a late answer but I have had the same problem and could just find one solution, there is an application that can update the model-first generated databases without losing the data.
It can directly open the model file and update the database tables.
It also installs some extensions on Visual Studio that I have not personally used but may be usable.
The name is Entity Developer and there are some editions of the application listed here:
Entity Developer Editions
The free edition is usable only for 10 entities or less that may not suit your needs but the Professional edition is usable for 30 day as a trial that may help you do the job. The only solution I could find on the net was this one.
Hope it helps you with the problem.