如何比較兩個表並選擇順序中的差異？

Question

我得到了兩張幾乎相同的表格，這兩張表格之間的某些行的順序是混合在一起的。我會給你一個例子，說明我的表看起來像：

table 1：

colum1 | colum2 | colum3 
-------+--------+----------- 
test1 | pizza | margarita 
test1 | pizza | cheese  
test1 | pizza | hawaii  
test1 | burger | salad  
test2 | pizza | margarita 
test2 | pizza | ham  
test2 | burger | tomat  
test2 | burger | salad  
test3 | pig | green  
test3 | cow | green  
test3 | cow | yellow  
test3 | pig | bird  

table 2：

colum1 | colum2 | colum3 
-------+--------+------------ 
test1 | pizza | margarita 
test1 | pizza | hawaii  <-- different spot than before 
test1 | pizza | cheese  <-- different spot than before 
test1 | burger | salad  
test2 | pizza | margarita 
test2 | pizza | ham  
------ | ------ | --------- <-- no value between those (so its missing) 
test2 | burger | salad  
test3 | pig | green  
test3 | cow | green  
test3 | cow | yellow  
test3 | pig | bird  

所以你可以在上面的表格中看到，有一排混合並且有一個缺失的行。現在我想進行如下查詢：「如果colum1相同，則檢查行是否與其他表的順序相同」如果不是這樣「將所有行顯示爲長如colum1具有相同的值「那麼，結果將是這樣的：

  table1     table2 
------+-------+-----------+-------+-------+----------- 
test1 | pizza | margarita | test1 | pizza | margarita 
test1 | pizza | cheese | test1 | pizza | hawaii  <<<shows all `test1` because this one is different 
test1 | pizza | hawaii | test1 | pizza | cheese  
test1 | burger| salad  | test1 | pizza | salad  
test2 | pizza | margarita | test2 | pizza | margarita 
test2 | pizza | ham  | test2 | pizza | ham  
test2 | burger| tomat  |  |  |    <<<shows all `test2` because this one is different 
test2 | burger| salad  | test2 | burger| salad  

另外：一行（這樣colum1，colum2，colum3組合的）始終相同。

那麼，這是SQL的某種可能嗎？或者我應該使用其他方法來做到這一點？

Answer 1

注意：您需要有一個索引，一個壞的ID生成它使用此代碼：

row_number() over (order by (select 0)) as rowId 

但是，如果你想跟蹤那些你可以使用查詢表之間的變化像這樣：

;with t1 as (
    select *, row_number() over (order by (select 0)) as rowId 
    from table1 
), t2 as (
    select *, row_number() over (order by (select 0)) as rowId 
    from table2 
), t as (
select t1.rowId, t1.colum1, t1.colum2, t1.colum3 
    , coalesce(t2.rowId - t1.rowId, 999999999) + sum(case when t2.rowId is null then 1 else 0 end) over (order by t1.rowId) cs 
from t1 full outer join t2 
    on t1.colum1 = t2.colum1 and t1.colum2 = t2.colum2 and t1.colum3 = t2.colum3 
) 
select t.colum1, t.colum2, t.colum3, 
    case when t.cs >= 999999999 then 'skipped' 
     when t.cs = 0 then '' 
     else 'replaced' 
    end table2State 
from t 
order by t.rowId; 

這導致這樣的：

colum1 | colum2 | colum3 | table2State 
--------+-----------+-----------+-------------- 
test1 | pizza  | margarita | 
test1 | pizza  | cheese | replaced 
test1 | pizza  | hawaii | replaced 
test1 | burger | salad  | 
test2 | pizza  | margarita | 
test2 | pizza  | ham  | 
test2 | burger | tomat  | skipped 
test2 | burger | salad  | 
test3 | pig  | green  | 

Answer 2

將帶有排序和row_number函數的表格與所有希望比較的列加上排序的數字進行左右連接，並且在表格不同的位置會出現空值。