We have a data API that has to publish data to an external system and keep the data up to date.
The external system needs aggregate/summary data from us.
We have ROWVERSION columns on all tables involved, but we need to create a view to publish the summary data.
Here is the core of the problem:
The parent table contains a ROWVERSION, lets call it Order. The child
table contains a ROWVERSION lets call it OrderItem. The VIEW needs to
publish Summary data for the Parent (Order) which includes columns
such as TotalAmount (Sum of OrderItem.Amounts for the Order) and
TotalItems (Count of OrderItems for the Order).
Now if we publish the ROWVERSION of the parent table then changes on
the child table(s) will not be reflected in the ROWVERSION.
So if we add or update OrderItems, the Order ROWVERSION will not
change but the OrderItem ROWVERSION will.
The External System needs to know that something changed on the Order,
but it cannot see the OrderItem table, and therefore can only look at
a ROWVERSION on the VIEW.
This cannot be a new problem and I'm sure there is a best practice design/solution to this problem.
I considered constructing a byte[] as ROWVERSION on the VIEW consisting of the sum of all of the other ROWVERSIONS of the Parent and Child table records. I'm unsure if this will work or what the possible issues could be. (Running out of number space as the rows increase for one...)
Any suggestions are welcome.
Related
I have a legacy database with a particular table -- I will call it ItemTable -- that can have billions of rows of data. To overcome database restrictions, we have decided to split the table into "silos" whenever the number of rows reaches 100,000,000. So, ItemTable will exist, then a procedure will run in the middle of the night to check the number of rows. If numberOfRows is > 100,000,000 then silo1_ItemTable will be created. Any Items added to the database from now on will be added to silo1_ItemTable (until it grows to big, then silo2_ItemTable will exist...)
ItemTable and silo1_ItemTable can be mapped to the same Item entity because the table structures are identical, but I am not sure how to set this mapping up at runtime, or how to specify the table name for my queries. All inserts should be added to the latest siloX_ItemTable, and all Reads should be from a specified siloX_ItemTable.
I have a separate siloTracker table that will give me the table name to insert/read the data from, but I am not sure how I can use this with entity framework...
Thoughts?
You could try to use the Entity Inheritance to get this. So you have a base class which has all the fields mapped to ItemTable and then you have descendant classes that inherit from ItemTable entity and is mapped to the silo tables in the db. Every time you create a new silo you create a new entity mapped to that silo table.
[Table("ItemTable")]
public class Item
{
//All the fields in the table goes here
}
[Table("silo1_ItemTable")]
public class Silo1Item : Item
{
}
[Table("silo2_ItemTable")]
public class Silo2Item : Item
{
}
You can find more information on this here
Other option is to create a view that creates a union of all those table and map your entity to that view.
As mentioned in my comment, to solve this problem I am using the SQLQuery method that is exposed by DBSet. Since all my item tables have the exact same schema, I can use the SQLQuery to define my own query and I can pass in the name of the table to the query. Tested on my system and it is working well.
See this link for an explanation of running raw queries with entity framework:
EF raw query documentation
If anyone has a better way to solve my question, please leave a comment.
[UPDATE]
I agree that stored procedures are also a great option, but for some reason my management is very resistant to make any changes to our database. It is easier for me (and our customers) to put the sql in code and acknowledge the fact that there is raw sql. At least I can hide it from the other layers rather easily.
[/UPDATE]
Possible solution for this problem may be using context initialization with DbCompiledModel param:
var builder = new DbModelBuilder(DbModelBuilderVersion.V6_0);
builder.Configurations.Add(new EntityTypeConfiguration<EntityName>());
builder.Entity<EntityName>().ToTable("TableNameDefinedInRuntime");
var dynamicContext = new MyDbContext(builder.Build(context.Database.Connection).Compile());
For some reason in EF6 it fails on second table request, but mapping inside context looks correct on the moment of execution.
We have a customer very large table with over 500 columns (i know someone does that!)
Many of these columns are in fact foreign keys to other tables.
We also have the requirement to eager load some of the related tables.
Is there any way in Linq to SQL or Dynamic Linq to specify what columns to be retrieved from the database?
I am looking for a linq statement that actually HAS this effect on the generated SQL Statement:
SELECT Id, Name FROM Book
When we run the reguar query generated by EF, SQL Server throws an error that you have reached the maximum number of columns that can be selected within a query!!!
Any help is much appreciated!
Yes exactly this is the case, the table has 500 columns and is self referencing our tool automatically eager loads the first level relations and this hits the SQL limit on number of columns that can be queried.
I was hoping that I can set to only load limited columns of the related Entities such as Id and Name (which is used in the UI to view the record to user)
I guess the other option is to control what FK columns should be eager loaded. However this still remains problem for tables that has a binary or ntext column which you may not want to load all the times.
Is there a way to hook multiple models (Entities) to the same table in Code First? We tried doing this I think the effort failed miserably.
Yes you can return only subset of columns by using projection:
var result = from x in context.LargeTable
select new { x.Id, x.Name };
The problem: projection and eager loading doesn't work together. Once you start using projections or custom joins you are changing shape of the query and you cannot use Include (EF will ignore it). The only way in such scenario is to manually include relations in the projected result set:
var result = from x in context.LargeTable
select new {
Id = x.Id,
Name = x.Name,
// You can filter or project relations as well
RelatedEnitites = x.SomeRelation.Where(...)
};
You can also project to specific type BUT that specific type must not be mapped (so you cannot for example project to LargeTable entity from my sample). Projection to the mapped entity can be done only on materialized data in Linq-to-objects.
Edit:
There is probably some misunderstanding how EF works. EF works on top of entities - entity is what you have mapped. If you map 500 columns to the entity, EF simply use that entity as you defined it. It means that querying loads entity and persisting saves entity.
Why it works this way? Entity is considered as atomic data structure and its data can be loaded and tracked only once - that is a key feature for ability to correctly persist changes back to the database. It doesn't mean that you should not load only subset of columns if you need it but you should understand that loading subset of columns doesn't define your original entity - it is considered as arbitrary view on data in your entity. This view is not tracked and cannot be persisted back to database without some additional effort (simply because EF doesn't hold any information about the origin of the projection).
EF also place some additional constraints on the ability to map the entity
Each table can be normally mapped only once. Why? Again because mapping table multiple times to different entities can break ability to correctly persist those entities - for example if any non-key column is mapped twice and you load instance of both entities mapped to the same record, which of mapped values will you use during saving changes?
There are two exceptions which allow you mapping table multiple times
Table per hierarchy inheritance - this is a mapping where table can contains records from multiple entity types defined in inheritance hierarchy. Columns mapped to the base entity in the hierarchy must be shared by all entities. Every derived entity type can have its own columns mapped to its specific properties (other entity types have these columns always empty). It is not possible to share column for derived properties among multiple entities. There must also be one additional column called discriminator telling EF which entity type is stored in the record - this columns cannot be mapped as property because it is already mapped as type discriminator.
Table splitting - this is direct solution for the single table mapping limitation. It allows you to split table into multiple entities with some constraints:
There must be one-to-one relation between entities. You have one central entity used to load the core data and all other entities are accessible through navigation properties from this entity. Eager loading, lazy loading and explicit loading works normally.
The relation is real 1-1 so both parts or relation must always exists.
Entities must not share any property except the key - this constraint will solve the initial problem because each modifiable property is mapped only once
Every entity from the split table must have a mapped key property
Insertion requires whole object graph to be populated because other entities can contain mapped required columns
Linq-to-Sql also contains ability to mark a column as lazy loaded but this feature is currently not available in EF - you can vote for that feature.
It leads to your options for optimization
Use projections to get read-only "view" for entity
You can do that in Linq query as I showed in the previous part of this answer
You can create database view and map it as a new "entity"
In EDMX you can also use Defining query or Query view to encapsulate either SQL or ESQL projection in your mapping
Use table splitting
EDMX allows you splitting table to many entities without any problem
Code first allows you splitting table as well but there are some problems when you split table to more than two entities (I think it requires each entity type to have navigation property to all other entity types from split table - that makes it really hard to use).
Create stored procedures that query the number of columns needed and then call the stored procs from code.
In our shop, when we design a database, we typically include auditing attributes for each table (LastUpdateUser, LastUpdateDate, etc). This is common practice, however, I've noticed this becoming an increasing problem when you have tables that "inherit" from other tables, especially using tools such as the entity framework.
For example, if you have tables Customers and Employees, and those tables have a foreign key to table People, then in your entity / class model when you establish the inheritance, you need to change the names for the audit fields because they exist in both tables. Perhaps they need to become PersonLastUpdatedUser and PersonLastUpdatedDate, while the ones from Employees remain as simply LastUpdatedUser and LastUpdatedDate.
When designing tables for inheritance, do you put such audit fields in both tables, or do you just have them in the parent table and update the parent table whenever an attribute changes in a child table?
If you want to use inheritance than those attributes belong to parent table because the parent with related table forms single entity and you track auditing for whole entity. If you for any reason needs those attributes in both tables it should be the first warning that those tables are not good candidates for inheritance.
If you want true auditing, you create separate audit tables that are populated by triggers (never ever by the application or you will miss items that need to be audited).
and they shouw both the old and new value as well as the date and the user or application that made the change.
If you want a last updatedcolumn in each table (which I think is better than having it only in the parenta as that doesn't tell you anything about which of the tables changes last) and you want o use inheritance then you might need to create unique names by adding the table name to lastUpdated. So PersonLastUpdated and OrderLastUpdated, etc.
Or you don't use inheritance.
I got one entity mapped to two tables. First of this tables has timestamp field.
What I want - if I modify field from first table it should check first table for concurrecny with timestamp column. The second table should'n updating. If I modify second table it should just update the second table. First table must be unchanged.
This is work if I set "Concurrecny=None" for the timestamp field in schema.
If I set "Concurrecny=Fixed" for the timestamp field and change properties from the second table it update the first table with current values.
How to make Concurrency check only for one of this two tables?
That is not possible. Once you map two tables to single entity they becomes one for entity framework and the time stamp is shared among them so if you do any changes to the second EF will always modify timestamp in the first. If you set Concurrency to None you are turning off the concurrency feature in EF and the main purpose of timestamp field.
I'm new to EF, and trying to work out the best way to do something
I have a procedure that returns the details of a table, but also a calculated value. What I'd like is for it to return this information to an entity that could contain this, whether it's the original entity, or an entity that's based on the original (so that if the original table changes, it will still map). What I'd like to avoid is maintaining 2 entities - 1 for the table, and 1 for the stored procedure results.
Thanks...
You could create a view that contained the table fields as well as the calculated values.