优化SQL查询，需要建议

Question

I have a table in SQL Server having 4 columns:

Invoice No, Date, Amt and ID

I have to find invoices that have same Invoice No, date and Amt but different ID. I'm populating the results doing self join but seems like it's not the optimized way to fetch results.

My query:

select * from table t1 join 
table t2 on t1.invoice = t2.invoice 
where t1.invoice=t2.invoice and t1.amount=t2.amount and t1.date =t2.date and t1.id!=t2.id

Kindly suggest me an optimized way to fetch the correct result.

Answer 1

try this. using left join and filter those nulls.

select * from (
    select t1.invoiceno, t1.date, t1.amt, t1.id, t2.id as t2ID
    from invoices t1
    left join invoices t2 on t2.invoiceno = t1.invoiceno 
        and t2.date = t1.date 
        and t2.amt = t1.amt
        and t2.id != t1.id) t3 
where coalesce(t3.t2ID, 0) != 0

Answer 2

You might use indexes to speed up the retrieving from large databases. Use sub query but don't use a sub query just to show one column.

I advised to use sub query as new table to use joins. just like the first answer.

Answer 3

use not exists

select t1.* from table t1 
where not exists( select 1 form 
table t2 where t1.invoice = t2.invoice 
and t1.invoice=t2.invoice and t1.amount=t2.amount 
and t1.date =t2.date and   t1.id=t2.id
having count(*)>1
)

Answer 4

have to find invoices that have same Invoice No, date and Amt but different ID.

Use exists :

select t.*
from t
where exists (select 1
              from t t2
              where t2.Invoice = t.invoice and
                    t2.Date = t.date and
                    t2.amount = t.amount and
                    t2.id <> t.id
             )
order by t.invoiceNo, t.date, t.amount, t.id;

This will show the matching invoices on adjacent rows. For performance, you want an index on (invoice, date, amount, id) .

If you just want triplets where this occurs, you can use aggregation:

select invoice, date, amount, min(id), max(id)
from t
group by invoice, date, amount
having count(distinct id) > 1;

Note: If there are more than two duplicates, this only shows two ids.