What's the recommended way of updating an entity? So far, I figured out two ways:
Just create a new entity with the existing Id and updated property values, and use session.SaveOrUpdate()
Use a DTO, retrieve the existing entity using session.Load(dto.Id), assign new vaues from the dto, then save.
No1 requires much less effort, but sometimes I'm getting an exception: "a different object with the same identifier value was already associated with the session". Is there a simple way around that?
No2 might require an extra trip to the DB I guess?
Sorry if that's been answered already, just couldn't find the answer.
Thanks
ulu
Your second option with DTOs is my preferred way. Your DTOs should be specific to the screen (Google Screen Bound DTOs) so that the screen and your domain can change independently of one another.
It also won't add an extra trip to the db since #1 would require a disconnected entity which would have to be reconnected (which triggers a select) after the fact. Worrying about one extra select also smells strongly of premature optimization.
In terms of converting from domain to DTO I'd recommend looking at AutoMapper.
To use No1 you could try and Evict the object from nHibernates session. This will get rid of the error about the object already being in the session.
I would recommend approach number 2. Especially if you want to add any sort of optomistic locking. In many cases a single extra hit to the db won't be that expensive.
Edit
To check if a entity already exists in the session you can use the Contains(obj) method on the Session instance.
Related
I thought Datastore's key was ordered by insertion date, but apparently I was wrong. I need to periodically look for new entities in the Datastore, fetch them and process them.
Until now, I would simply store the last fetched key and wrongly query for anything greater than it.
Is there a way of doing so?
Thanks in advance.
Datastore automatically generated keys are generated with uniform distribution, in order to make search more performant. You will not be able to understand which entity where added last using keys.
Instead, you can try couple of different approaches.
Use Pub/Sub and architecture your app so another background task will consume this last added entities. On entities add in DB, you will just publish new Event into Pub/Sub with key id. You event listener (separate routine) will receive it.
Use names and generate you custom names. But, as you want to create sequentially growing names, this will case performance hit on even not big ranges of data. You can find more about this in Best Practices of Google Datastore.
https://cloud.google.com/datastore/docs/best-practices#keys
You can add additional creation time column, and still use automatic keys generation.
I am writing a web application that allows the user basic CRUD operations against a database. The tables that are being updated have less than 200 records and there may be multiple users using this applications there is a need for some sort of locking mechanism to avoid the 2 users from overwriting each others changes.
I have looked into semaphores but that seems to only limit the number of users executing the same code. In my data layer I have a class file for each table so I can certainly employ this on a specific table's class file but can I somehow limit the locking to the key fields?
Assuming that you are using a proper SQL implementation along with ASP .Net, why dont you use transactions to achieve this? Check it out here.
Additionally, you can also read up on optimistic concurrency to see if that is what you need. Basically, before saving a value, the user checks if the value in a particular field is the same as it was when he first read it. If the value is the same, it is assumed that noone else has overwritten it, and the new value is saved to the DB; if the values are not the same, a warning message is returned instead.
I have seen a particular pattern a few times over the last few years. Please let me describe it.
In the UI, each new record (e.g., new customers details) is stored on the form without saving to database. This clearly has been done so not clutter the database or cause unnecessary database hits.
While in the UI state, these objects are identified using a Guid. When these are a saved to the database, their associated Guids are not stored. Instead, they are assigned a database Int as their primary key.
The form can cope with a mixure of retrieved items from the database (using Int) as well as those that have not yet been committed (using Guid).
When inspecting the form (using Firebug) to see which key was used, we found a two part delimited combined key had been used. The first part is a guid (an empty guid if drawn from the database) and the second part is the integer (zero is stored if it is not drawn from the database). As one part of the combined key will always uniquely identify a record, it works rather well.
Is this Good practice or not? Can anyone tell me the pattern name or suggest one if it is not already named?
There are a couple patterns at play here.
Identity Field Pattern
Defined in P of EAA as "Saves a database ID field in an object to maintain identity between an in-memory object and a database row." This part is obvious.
Transaction Script and Metadata Mapping
In general, the ASP.NET DataBound controls use something like an Transaction Script pattern in conjunction with a Metadata Mapping pattern. Fowler defines Metadata Mapping as "holding details of object-relational mapping in metadata". If you have ever written a data source control, the Metadata Mapping aspect of this pattern seems obvious.
The Transaction Script pattern "organizes business logic by procedures where each procedure handles a single request from the presentation." In order to encapsulate the logic of maintaining both presentation state and data-state it is necessary for the intermediary object to indicate:
If a database record exists
How to identify the backend data record, to populate the UI control
How to identify the data and the UI control if there is no current data record, so that presentation data can be updated from the backend datastore.
The presence of the new client data entry Guid and the data-record integer Id provide adequate information to determine all of this with only a single call to the database. This could be accomplished by just using integers (and perhaps giving a unique negative integer for each unpersisted UI data item), but it is probably more explicit to have two separate fields.
Good or Bad Practice?
It depends. ASP.NET is a pretty successful software project, and this pattern seems to work consistently. However, this type of ASP.NET web control has a very specific scope of application - to encapsulate interaction between a UI and a database about data objects with simple mappings. The concerns do seem a little blurred, but for many applicable scenarios this will still be acceptable. The pattern is valid whereever a Row Data Gateway would be acceptable. If there is more than one database row affected by a web control, then this approach will not be functional. In these more complex cases, either an Active Record implementation or the combination of a Domain Model and a Repository implementation would be better suited.
Whether a pattern is good or bad practice really depends on the scenario in which it is being applied. It seems like people tend to advocate more complex design structures, because they can be applied to more scenarios without failing. However, in a very simple application where the mappings between data records and the UI are direct, this pattern is very useful because it creates the intended result while minimizing the amount of performance and development overhead.
I don't think there is a specific pattern for that.
Is it good practice? I don't think so. First, it's not very object oriented. How about:
interface ICommittable
{
/// <summary>
/// Gets or sets a value indicating whether the entity was already committed to the database.
/// </summary>
bool IsCommitted;
/// <summary>
/// Gets or sets the ID of the entity, used either in database or generated by UI or an underlying BL.
/// </summary>
Guid Id;
}
Instead, what they do is to mix three separate data entries in one in a non obvious way:
The ID
Another ID (why?)
A fact that the entity was committed or not.
Especially, having two separate IDs is extremely confusing and will require not only a good documentation, but some time for a new developer to understand what's happening here.
If the purpose was to create new entities without querying the database for a new ID, they could use GUIDs everywhere: when a new entity is created, you Guid.CreateNew it's ID, then, if need, you commit everything, this GUID being the identifier in the database too (there are few chances to have a collision between already saved GUIDs and a new one, so I wouldn't care about that).
Much more simple, isn't it?
It's also not easy to do a few things. For example, how do you compare two entities? Remember that:
Two committed entities which have different GUIDs are not equal,
Two not committed entities which have different IDs are not equal,
A committed entity may be equal to a non committed entity, even if their GUIDs and their IDs will be different.
To conclude, it seems like a lack of refactoring. Probably they were modifying a project where entities were already identified in the database by their id (int) unique key, so instead of refactoring this, they just added GUIDs, thus making the overall thing:
More difficult to understand,
Very difficult to work with and to modify in future.
If I'm not wrong it's the repository pattern: http://martinfowler.com/eaaCatalog/repository.html
This is well described in the Evans Domain Driven Design book and has proven to work well under specific circumstances.
I have asked a few questions today as I try to think through to the solution of a problem.
We have a complex data structure where all of the various entities are tightly interconnected, with almost all entities heavily reliant/dependant upon entities of other types.
The project is a website (MVC3, .NET 4), and all of the logic is implemented using LINQ-to-SQL (2008) in the business layer.
What we need to do is have a user "lock" the system while they make their changes (there are other reasons for this which I won't go into here that are not database related). While this user is making their changes we want to be able to show them the original state of entities which they are updating, as well as a "preview" of the changes they have made. When finished, they need to be able to rollback/commit.
We have considered these options:
Holding open a transaction for the length of time a user takes to make multiple changes stinks, so that's out.
Holding a copy of all the data in memory (or cached to disk) is an option but there is heck of a lot of it, so seems unreasonable.
Maintaining a set of secondary tables, or attempting to use session state to store changes, but this is complex and difficult to maintain.
Using two databases, flipping between them by connection string, and using T-SQL to manage replication, putting them back in sync after commit/rollback. I.e. switching on/off, forcing snapshot, reversing direction etc.
We're a bit stumped for a solution that is relatively easy to maintain. Any suggestions?
Our solution to a similar problem is to use a locking table that holds locks per entity type in our system. When the client application wants to edit an entity, we do a "GetWithLock" which gets the client the most up-to-date version of the entity's data as well as obtaining a lock (a GUID that is stored in the lock table along with the entity type and the entity ID). This prevents other users from editing the same entity. When you commit your changes with an update, you release the lock by deleting the lock record from the lock table. Since stored procedures are the api we use for interacting with the database, this allows a very straight forward way to lock/unlock access to specific entities.
On the client side, we implement IEditableObject on the UI model classes. Our model classes hold a reference to the instance of the service entity that was retrieved on the service call. This allows the UI to do a Begin/End/Cancel Edit and do the commit or rollback as necessary. By holding the instance of the original service entity, we are able to see the original and current data, which would allow the user to get that "preview" you're looking for.
While our solution does not implement LINQ, I don't believe there's anything unique in our approach that would prevent you from using LINQ as well.
HTH
Consider this:
Long transactions makes system less scalable. If you do UPDATE command, update locks last until commit/rollback, preventing other transaction to proceed.
Second tables/database can be modified by concurent transactions, so you cannot rely on data in tables. Only way is to lock it => see no1.
Serializable transaction in some data engines uses versions of data in your tables. So after first cmd is executed, transaction can see exact data available in cmd execution time. This might help you to show changes made by user, but you have no guarantee to save them back into storage.
DataSets contains old/new version of data. But that is unfortunatelly out of your technology aim.
Use a set of secondary tables.
The problem is that your connection should see two versions of data while the other connections should see only one (or two, one of them being their own).
While it is possible theoretically and is implemented in Oracle using flashbacks, SQL Server does not support it natively, since it has no means to query previous versions of the records.
You can issue a query like this:
SELECT *
FROM mytable
AS OF TIMESTAMP
TO_TIMESTAMP('2010-01-17')
in Oracle but not in SQL Server.
This means that you need to implement this functionality yourself (placing the new versions of rows into your own tables).
Sounds like an ugly problem, and raises a whole lot of questions you won't be able to go into on SO. I got the following idea while reading your problem, and while it "smells" as bad as the others you list, it may help you work up an eventual solution.
First, have some kind of locking system, as described by #user580122, to flag/record the fact that one of these transactions is going on. (Be sure to include some kind of periodic automated check, to test for lost or abandoned transactions!)
Next, for every change you make to the database, log it somehow, either in the application or in a dedicated table somewhere. The idea is, given a copy of the database at state X, you could re-run the steps submitted by the user at any time.
Next up is figuring out how to use database snapshots. Read up on these in BOL; the general idea is you create a point-in-time snapshot of the database, do whatever you want with it, and eventually throw it away. (Only available in SQL 2005 and up, Enterprise edition only.)
So:
A user comes along and initiates one of these meta-transactions.
A flag is marked in the database showing what is going on. A new transaction cannot be started if one is already in process. (Again, check for lost transactions now and then!)
Every change made to the database is tracked and recorded in such a fashion that it could be repeated.
If the user decides to cancel the transaction, you just drop the snapshot, and nothing is changed.
If the user decides to keep the transaction, you drop the snapshot, and then immediately re-apply the logged changes to the "real" database. This should work, since your requirements imply that, while someone is working on one of these, no one else can touch the related parts of the database.
Yep, this sure smells, and it may not apply to well to your problem. Hopefully the ideas here help you work something out.
I need to manage the acquisition of many record at hour. About 1000000 records. And I need to get every second the last insert value for every primary key. It works quit well with sharding. I was thinking to try the use os capped collection to get only the last record for every primary key. In order to do this, I made two separated insert, there is a way, into mongodb, to make some kind of trigger to propagate the insert into a collection to another collection?
MongoDB does not have any support for triggers or similar behavior.
The only way to do this is to make it happen in your code. So the code that writes the first entry should also write the second.
People have definitely requested triggers. If they are necessary for your solution, please cast a vote on the feature request.
I disagree with "triggers is needed". People, MongoDB was created to be very fast and to provide as basic functionalities as can be. This is a power of this solution.
I think that here the best think is to create triggers inside Your application as a part of Data Access layer.