LINQ: combining join and group by

I met the same problem as you.

I push two tables result into t1 object and group t1.

 from p in Products                         
  join bp in BaseProducts on p.BaseProductId equals bp.Id
  select new {
   p,
   bp
  } into t1
 group t1 by t1.p.SomeId into g
 select new ProductPriceMinMax { 
  SomeId = g.FirstOrDefault().p.SomeId, 
  CountryCode = g.FirstOrDefault().p.CountryCode, 
  MinPrice = g.Min(m => m.bp.Price), 
  MaxPrice = g.Max(m => m.bp.Price),
  BaseProductName = g.FirstOrDefault().bp.Name
};

Once you've done this

group p by p.SomeId into pg  

you no longer have access to the range variables used in the initial from. That is, you can no longer talk about p or bp, you can only talk about pg.

Now, pg is a group and so contains more than one product. All the products in a given pg group have the same SomeId (since that's what you grouped by), but I don't know if that means they all have the same BaseProductId.

To get a base product name, you have to pick a particular product in the pg group (As you are doing with SomeId and CountryCode), and then join to BaseProducts.

var result = from p in Products                         
 group p by p.SomeId into pg                         
 // join *after* group
 join bp in BaseProducts on pg.FirstOrDefault().BaseProductId equals bp.Id         
 select new ProductPriceMinMax { 
       SomeId = pg.FirstOrDefault().SomeId, 
       CountryCode = pg.FirstOrDefault().CountryCode, 
       MinPrice = pg.Min(m => m.Price), 
       MaxPrice = pg.Max(m => m.Price),
       BaseProductName = bp.Name  // now there is a 'bp' in scope
 };

That said, this looks pretty unusual and I think you should step back and consider what you are actually trying to retrieve.


We did it like this:

from p in Products                         
join bp in BaseProducts on p.BaseProductId equals bp.Id                    
where !string.IsNullOrEmpty(p.SomeId) && p.LastPublished >= lastDate                         
group new { p, bp } by new { p.SomeId } into pg    
let firstproductgroup = pg.FirstOrDefault()
let product = firstproductgroup.p
let baseproduct = firstproductgroup.bp
let minprice = pg.Min(m => m.p.Price)
let maxprice = pg.Max(m => m.p.Price)
select new ProductPriceMinMax
{
SomeId = product.SomeId,
BaseProductName = baseproduct.Name,
CountryCode = product.CountryCode,
MinPrice = minprice, 
MaxPrice = maxprice
};

EDIT: we used the version of AakashM, because it has better performance