SQL or LINQ for this functionality?

I'm writing an SQL script to update the database of a deployed application (add a few tables and copy/update some data) to accommodate new functionality being added to the application (.NET using C#). I've written some code in the application to handle such updates (whenever a new update is available, run its .sql) which reusable for all DB updates (simply provide a new .sql).

The problem I'm running into now is that I am fairly new to SQL, and have a pretty short timeline to get this update ready. The more complex part of the update could easily be written in the application, but for consistency's sake I'd like to do it in SQL.

What I need to do, I've summarized as the following algorithm/pseudocode:

for each category in allCategories
    Select * from unrankedEntities
    where Category = category
    order by (score)
    take top Category.cutoff
    insert each of these into rankedEntities with rank = 0

    Select * from unrankedEntities
    where Category = category
    order by (score) DESC
    take top 16 - Category.cutoff
    insert each of these into rankedEntities wit开发者_C百科h rank = null, belowcutoff=true

This is the general algorithm I am trying to implement (with a couple more tweaks to be made after this part is complete). Would it be more efficient to write this in SQL or just write the LINQ to do it (resulting in changing the application's database update code for each DB update)? If SQL is the best choice, where can I find a good reference/tutorial covering the topics I would need to do this (I've looked through few different ones online, and some seem pretty simple, so I would appreciate a recommendation of a good one).

Additionally, if you have any tips on improving the algorithm, I'd be happy to hear your advice.

EDIT:

I forgot to mention, but (score) is not a value stored in unrankedEntities. It needs to be calclulated for each entity by counting rows in another table that match a specific condition, and mulitplying by a constant and by numbers pulled from other tables. Additionally, it needs to be stored into rankedEntities to avoid all this calculation every time it is needed (often).

EDIT:

It has been pointed out to me that parts of the new additions to the application already contain the code necessary to do all of the data manipulation, so I will go with the simple solution of recycling that code, although the result is a less tidy update-via-sql class, as it now has an update-via-sql section and update-via-code section.

When using LINQ to SQL in debug mode you can see the SQL it's generating, which you could copy into a script, if you need to hurry and not learn all about SQL right at this moment. See http://weblogs.asp.net/scottgu/archive/2007/07/31/linq-to-sql-debug-visualizer.aspx

EDIT: I didn't realize you were using Entity Framework. See this response to a similar question: How do I get the raw SQL underlying a LINQ query when using Entity Framework CTP 5 "code only"?.

LINQ is perfectly fine for this, but inserts and updates each entity in a singular fashion, meaning in SQL you could update in bulk with one query, but you cannot in LINQ... it generates separate statements for each entity. So SQL has some efficiency gains. If you are more comfortable with LINQ, do it in LINQ. If you are talking thousands of objects, you may want to turn off change tracking (turned off in EF by changing the merge option) as the internal memory of changed objects will grow rather large.