LINQ-如何使用按顺序排序？

Question

I want to do a distinct on this:

query = from acao in query
  join itemAuditoria in Session.Query<ItemAuditoria>() on acao.Id equals itemAuditoria.Acao.Id
  join auditoria in Session.Query<Auditoria>() on itemAuditoria.Auditoria.Id equals auditoria.Id
  join maquina in Session.Query<Maquina>() on auditoria.Maquina.Id equals maquina.Id
  orderby maquina.Nome, acao.Numero
  select acao;

But if I try putting a distinct at the end of the query like this:

query = (from acao in query
  join itemAuditoria in Session.Query<ItemAuditoria>() on acao.Id equals itemAuditoria.Acao.Id
  join auditoria in Session.Query<Auditoria>() on itemAuditoria.Auditoria.Id equals auditoria.Id
  join maquina in Session.Query<Maquina>() on auditoria.Maquina.Id equals maquina.Id
  orderby maquina.Nome, acao.Numero
  select acao).Distinct();

then I get this error:

ORDER BY items must appear in the select list if SELECT DISTINCT is specified.

Answer 1

Not sure about the performance impact compared to the answer in OP's comments, but this could be subqueried for readability!

query =
    (from distinctAcao in
        (from acao in query
        join itemAuditoria in Session.Query<ItemAuditoria>() on acao.Id equals itemAuditoria.Acao.Id
        join auditoria in Session.Query<Auditoria>() on itemAuditoria.Auditoria.Id equals auditoria.Id
        join maquina in Session.Query<Maquina>() on auditoria.Maquina.Id equals maquina.Id
        select acao).Distinct()
    orderby maquina.Nome, distinctAcao.Numero
    select distinctAcao);

---

edit:

So sorry for overlooking that it didn't compile - it did in my head ;) Corrections on the way!

Because we're doing subquery here, the result of which is a list of acao items which don't have a Nome field, we cannot sort by it after we leave the subquery. Therefore we can either:

1. Sort inside subquery, then eliminate duplicates.

    query=( from distinctAcao in ( from acao in query join itemAuditoria in Session.Query<ItemAuditoria>() on acao.Id equals itemAuditoria.Acao.Id join auditoria in Session.Query<Auditoria>() on itemAuditoria.Auditoria.Id equals auditoria.Id join maquina in Session.Query<Maquina>() on auditoria.Maquina.Id equals maquina.Id orderby maquina.Nome,acao.Numero select acao ).Distinct() select distinctAcao ); 

   Result: sorted, unique acao s list

   Cons: first sorting, then filtering might impact performance

2. Return a better, combined object from the subquery, which will expose new fields from joined tables.

    query2=( from distinctMix in ( from acao in query join itemAuditoria in Session.Query<ItemAuditoria>() on acao.Id equals itemAuditoria.Acao.Id join auditoria in Session.Query<Auditoria>() on itemAuditoria.Auditoria.Id equals auditoria.Id join maquina in Session.Query<Maquina>() on auditoria.Maquina.Id equals maquina.Id select new { Id=acao.Id,Numero=acao.Numero,NomeFromMaquina=maquina.Nome } ).Distinct() orderby distinctMix.NomeFromMaquina,distinctMix.Numero select distinctMix ); 

   Result: quickly sorted, unique ( acao +more) list

   Cons: returned object will be of different type than original query that was used to iterate acao s, so assigning query=from acao in query select new { /* something else than acao */ }; won't work (hence query2 in this code example)

3. Use answer from OP's comments :)

PS

This time the code compiles for sure - test code available at https://ideone.com/mDRRle :>