linq 中的子查询和分组依据

Question

这个查询是不久前在我们的系统中编写的，但是随着数据的一点点增加，这个查询的性能变得很差。我的调查显示 (CodeCount) 查询触发另一个子查询导致执行的大量延迟。我需要优化这个 Linq 查询。任何帮助将不胜感激

  from batch in Context.VoucherCodeBatch.ToList()
                    join type in Context.VoucherCodeType on batch.VoucherTypeId equals type.VoucherTypeId
                    join voucher in Context.Voucher on batch.VoucherCodeBatchId equals voucher.VoucherCodeBatchId

                    where batchIds.Contains(batch.BatchCode)
                    group new
                    {
                        batch.BatchCode,
                        batch.CreationDate,
                        type.VoucherTypeName,
                        voucher.AllowedCount,
                        voucher.ValidFrom,
                        voucher.ValidTo,
                        batch.VoucherCodeBatchId,
                        voucher.VoucherCode
                    }
                        by new { batch.BatchCode }
                        into uniquebatch
                        select new Batch
                        {
                            BatchCode = uniquebatch.FirstOrDefault().BatchCode,
                            CreationDate = uniquebatch.FirstOrDefault().CreationDate,
                            TimesAllowed = uniquebatch.FirstOrDefault().AllowedCount,
                            ValidFrom = uniquebatch.FirstOrDefault().ValidFrom,
                            CodeCount = ((from c in Context.Voucher.ToList()
                                          where
                                              c.VoucherCodeBatchId ==
                                              uniquebatch.FirstOrDefault().VoucherCodeBatchId
                                          select c).Count()),
                            ValidTo = uniquebatch.FirstOrDefault().ValidTo,
                            CodeType = uniquebatch.FirstOrDefault().VoucherTypeName,
                            VoucherCodeBatchId = uniquebatch.FirstOrDefault().VoucherCodeBatchId
                        });

Answer 1

第一个大问题是 ObjectSet（EF 中的表集合）前面的 ToList()。

永远不要这样做，ToList() 会强制 EF 在处理查询之前将所有数据带入内存。 （喜欢@Daniel Hilgarth 的评论）。

其他细节是在FirstOrDefault() 前面使用get 属性，如行中：

BatchCode = uniquebatch.FirstOrDefault().BatchCode,

在这种情况下使用First() 代替FirstOrDefault。喜欢：

BatchCode = uniquebatch.First().BatchCode,

您的查询将是这样的：

from batch in Context.VoucherCodeBatch/*.ToList()*/
join type in Context.VoucherCodeType on batch.VoucherTypeId equals type.VoucherTypeId
join voucher in Context.Voucher on batch.VoucherCodeBatchId equals voucher.VoucherCodeBatchId
where batchIds.Contains(batch.BatchCode)
group new
    {
        batch.BatchCode,
        batch.CreationDate,
        type.VoucherTypeName,
        voucher.AllowedCount,
        voucher.ValidFrom,
        voucher.ValidTo,
        batch.VoucherCodeBatchId,
        voucher.VoucherCode
    }
by new { batch.BatchCode }
into uniquebatch
select ( delegate
    {
        // If you put a operation in a query that operation will be
        // processed all times. Bacause that i removed this line from
        // the where statement.
        var vcBatchId = uniquebatch.First().VoucherCodeBatchId;

        return new Batch
            {
                BatchCode = uniquebatch.First().BatchCode,
                CreationDate = uniquebatch.First().CreationDate,
                TimesAllowed = uniquebatch.First().AllowedCount,
                ValidFrom = uniquebatch.First().ValidFrom,
                CodeCount = ((
                    from c in Context.Voucher/*.ToList()*/
                    where c.VoucherCodeBatchId == vcBatchId
                    select c).Count()),
                ValidTo = uniquebatch.First().ValidTo,
                CodeType = uniquebatch.First().VoucherTypeName,
                VoucherCodeBatchId = uniquebatch.First().VoucherCodeBatchId
            }
    });

如果这个改进不够好，您需要将此查询更改为 SQL。但我相信这种改进会表现得更好。