Question

I need to update a column in a table which contains a lot of rows. Each row has a some large TEXT columns in it, which i do not need for my update.

I’m using LinqPAD, and this is roughly, what i wanna do:

(from s in Table
where s.FK_ID == null 
select new{s.FK_ID, s.Datum, s.PBNummer}).ToList()

.ForEach(s => s.FK_ID = new Guid(...some new guid here...));

SubmitChanges();

This does not compile, as the properties of an anonymous class type are read-only.

If I do

(from s in Table
where s.FK_ID == null 
select s).ToList()

then I can update and save, but all columns are loaded, which takes a very long time and causes memory problems.

Is there a way to only load some columns but still have an object that i can update and save using SubmitChanges? Or do i have to switch to SQL statements?

Answer 1

Firstly, if you don’t have a primary key in the database, then you wouldn’t be able to update via Linq-To-Sql. If you have a primary key, but just don’t know which it is, you can find it in Linqpad by doing something like

var table =  (from t in Mapping.GetTables() 
              where t.TableName == "[Table]" select t).SingleOrDefault();

(from dm in table.RowType.DataMembers 
           where dm.DbType != null && dm.IsPrimaryKey 
           select dm.Name)
          .Dump("Primary Key");

Once you know the primary key, you can do something like the following, (I’m assuming the primary key is called Id)

var oldList = (from s in Table
          where s.FK_ID == null 
          select new{s.Id , s.FK_ID, s.Datum, s.PBNummer}).ToList() ;

This is similar to your query, except I have added the primary key

foreach(var r in oldList)
{
    Table t = new Table();
    t.Id        = r.Id ;

    Table.Attach(t);
    t.FK_ID   = new Guid(...some new guid here...));
}
SubmitChanges();