T-SQLを使用して列の値に基づいて行のランク付けを行う方法は？

Question

編集：

私は、このコードを選択された溶液を改変し、それが意図した結果を達成しました。私は別の列の値に応じて行にランク値を割り当てるに行くかどう

DENSE_RANK() 
    OVER (ORDER BY 
     CASE WHEN f.empl_ssn = f.cov_ind_ssn 
     THEN 0 ELSE 1 END, f.cov_ind_ssn) AS [depend_no] 

---

？私は、RANK（）、ROW_NUMBER（）、DENSE_RANK（）関数を使用して読み込みました。特定の状況でDENSE_RANK（）を動作させてテストを行いましたが、1家族から172,000行その解決策が失敗します。私はこのようなルックスを達成するために必要なもの

：

+-------------+-------------+--------+--------+------+---------------+-----------+ 
| empl_ssn | cov_ind_ssn | f_name | l_name | code | date_of_birth | depend_no | 
+-------------+-------------+--------+--------+------+---------------+-----------+ 
| 000-11-0010 | 000-11-0011 | Joe | Bob | 1 | 12/25/72  |   2 | 
| 000-11-0010 | 000-11-0010 | Morgan | Bob | 18 | 9/14/71  |   1 | 
| 000-11-0010 | 000-12-0102 | Little | Bob | 19 | 2/16/06  |   3 | 
| 000-11-0010 | 000-51-9212 | Wittle | Bob | 19 | 8/12/04  |   4 | 
| 000-11-0010 | 000-53-0013 | Tiny | Bob | 19 | 8/25/02  |   5 | 
+-------------+-------------+--------+--------+------+---------------+-----------+ 

このデータは、任意の順序になりますが、私はすでに1に人ダウンあたり12行からデータを取るために、いくつかのCTEのを使用しました行ごとに、私は今それらにdepend_noを割り当てる必要があります。

empl_ssnとcov_ind_ssnが等しい場合、depend_noは1である必要があります。これは、それらが会社の従業員であることを意味します。それ以降は2から始まり、すべての行に値が割り当てられるまでインクリメントする必要があります。

---

は今、私はこのCTEを使用することによって、これを達成した：

random AS 
(
    SELECT DISTINCT 
    empl_ssn, 
    cov_ind_ssn, 
    relationship_code, 
    DENSE_RANK() 
    OVER (
     ORDER BY f.birthdate) AS [depend_no] 
    FROM full_data AS f 
    WHERE empl_ssn = @current_employee 
) 

しかし、私は従業員が常に最初に生まれてきたことを保証することはできませんので、それは明らかに実行可能ではありません。

ちょうど参考のために、これまでに書いたブロック全体を、慎重にストライプ化して投稿します。

DECLARE @current_employee VARCHAR(11); 
SET @current_employee = 'XXX-XX-XXXX'; 

WITH premiums AS 
(
    SELECT * 
    FROM 
     (
     SELECT 
      empl_ssn, 
      premium_amount, 
      month_covered 
     FROM full_data 
     WHERE empl_ssn = @current_employee 
    ) AS src 
     PIVOT (
     MAX(premium_amount) 
     FOR month_covered 
     IN ([1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12])) AS piv1 
), 
    offer_codes AS 
    (
     SELECT * 
     FROM 
     (
      SELECT 
      empl_ssn, 
      offer_code, 
      month_covered 
      FROM full_data 
      WHERE empl_ssn = @current_employee 
     ) AS src 
     PIVOT (
      MAX(offer_code) 
     FOR month_covered 
     IN ([1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12])) AS piv1 
), 
    safe_harbor_codes AS 
    (
     SELECT * 
     FROM 
     (
      SELECT 
      empl_ssn, 
      safe_harbor_code, 
      month_covered 
      FROM full_data 
      WHERE empl_ssn = @current_employee 
     ) AS src 
     PIVOT (
      MAX(safe_harbor_code) 
     FOR month_covered 
     IN ([1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12])) AS piv1 
), 
    random AS 
    (
     SELECT DISTINCT 
     empl_ssn, 
     cov_ind_ssn, 
     relationship_code, 
     DENSE_RANK() 
     OVER (
      ORDER BY f.birthdate) AS [depend_no] 
     FROM full_data AS f 
     WHERE empl_ssn = @current_employee 
) 

SELECT DISTINCT 
    f.empl_ssn, 
    f.cov_ind_ssn, 
    f.f_name, 
    f.l_name, 
    f.relationship_code, 
    f.birthdate, 

    r.depend_no, 

    CASE WHEN p.[1] = p.[12] 
    THEN NULL 
    ELSE 
    p.[1] END AS [emp_share_jan], 
    CASE WHEN p.[1] = p.[12] 
    THEN NULL 
    ELSE 
    p.[2] END AS [emp_share_feb], 
    CASE WHEN p.[1] = p.[12] 
    THEN NULL 
    ELSE 
    p.[3] END AS [emp_share_mar], 
    CASE WHEN p.[1] = p.[12] 
    THEN NULL 
    ELSE 
    p.[4] END AS [emp_share_apr], 
    CASE WHEN p.[1] = p.[12] 
    THEN NULL 
    ELSE 
    p.[5] END AS [emp_share_may], 
    CASE WHEN p.[1] = p.[12] 
    THEN NULL 
    ELSE 
    p.[6] END AS [emp_share_jun], 
    CASE WHEN p.[1] = p.[12] 
    THEN NULL 
    ELSE 
    p.[7] END AS [emp_share_jul], 
    CASE WHEN p.[1] = p.[12] 
    THEN NULL 
    ELSE 
    p.[8] END AS [emp_share_aug], 
    CASE WHEN p.[1] = p.[12] 
    THEN NULL 
    ELSE 
    p.[9] END AS [emp_share_sep], 
    CASE WHEN p.[1] = p.[12] 
    THEN NULL 
    ELSE 
    p.[10] END AS [emp_share_oct], 
    CASE WHEN p.[1] = p.[12] 
    THEN NULL 
    ELSE 
    p.[11] END AS [emp_share_nov], 
    CASE WHEN p.[1] = p.[12] 
    THEN NULL 
    ELSE 
    p.[12] END AS [emp_share_dec], 

    CASE WHEN o.[1] = o.[12] 
    THEN NULL 
    ELSE 
    o.[1] END AS [offer_code_jan], 
    CASE WHEN o.[1] = o.[12] 
    THEN NULL 
    ELSE 
    o.[2] END AS [offer_code_feb], 
    CASE WHEN o.[1] = o.[12] 
    THEN NULL 
    ELSE 
    o.[3] END AS [offer_code_mar], 
    CASE WHEN o.[1] = o.[12] 
    THEN NULL 
    ELSE 
    o.[4] END AS [offer_code_apr], 
    CASE WHEN o.[1] = o.[12] 
    THEN NULL 
    ELSE 
    o.[5] END AS [offer_code_may], 
    CASE WHEN o.[1] = o.[12] 
    THEN NULL 
    ELSE 
    o.[6] END AS [offer_code_jun], 
    CASE WHEN o.[1] = o.[12] 
    THEN NULL 
    ELSE 
    o.[7] END AS [offer_code_jul], 
    CASE WHEN o.[1] = o.[12] 
    THEN NULL 
    ELSE 
    o.[8] END AS [offer_code_aug], 
    CASE WHEN o.[1] = o.[12] 
    THEN NULL 
    ELSE 
    o.[9] END AS [offer_code_sep], 
    CASE WHEN o.[1] = o.[12] 
    THEN NULL 
    ELSE 
    o.[10] END AS [offer_code_oct], 
    CASE WHEN o.[1] = o.[12] 
    THEN NULL 
    ELSE 
    o.[11] END AS [offer_code_nov], 
    CASE WHEN o.[1] = o.[12] 
    THEN NULL 
    ELSE 
    o.[12] END AS [offer_code_dec], 

    CASE WHEN S.[1] = S.[12] 
    THEN NULL 
    ELSE 
    S.[1] END AS [safe_harbor_jan], 
    CASE WHEN S.[1] = S.[12] 
    THEN NULL 
    ELSE 
    S.[2] END AS [safe_harbor_feb], 
    CASE WHEN S.[1] = S.[12] 
    THEN NULL 
    ELSE 
    S.[3] END AS [safe_harbor_mar], 
    CASE WHEN S.[1] = S.[12] 
    THEN NULL 
    ELSE 
    S.[4] END AS [safe_harbor_apr], 
    CASE WHEN S.[1] = S.[12] 
    THEN NULL 
    ELSE 
    S.[5] END AS [safe_harbor_may], 
    CASE WHEN S.[1] = S.[12] 
    THEN NULL 
    ELSE 
    S.[6] END AS [safe_harbor_jun], 
    CASE WHEN S.[1] = S.[12] 
    THEN NULL 
    ELSE 
    S.[7] END AS [safe_harbor_jul], 
    CASE WHEN S.[1] = S.[12] 
    THEN NULL 
    ELSE 
    S.[8] END AS [safe_harbor_aug], 
    CASE WHEN S.[1] = S.[12] 
    THEN NULL 
    ELSE 
    S.[9] END AS [safe_harbor_sep], 
    CASE WHEN S.[1] = S.[12] 
    THEN NULL 
    ELSE 
    S.[10] END AS [safe_harbor_oct], 
    CASE WHEN S.[1] = S.[12] 
    THEN NULL 
    ELSE 
    S.[11] END AS [safe_harbor_nov], 
    CASE WHEN S.[1] = S.[12] 
    THEN NULL 
    ELSE 
    S.[12] END AS [safe_harbor_dec], 

    CASE WHEN p.[1] = p.[12] 
    THEN p.[1] 
    ELSE NULL END AS [premium_all], 
    CASE WHEN o.[1] <> o.[12] 
    THEN NULL 
    ELSE o.[1] END AS [offer_all], 
    CASE WHEN S.[1] <> S.[12] 
    THEN NULL 
    ELSE S.[1] END AS [harbor_all] 
FROM 
    premiums AS p 
    JOIN offer_codes AS o ON p.empl_ssn = o.empl_ssn 
    JOIN safe_harbor_codes AS S ON p.empl_ssn = S.empl_ssn 
    JOIN full_data AS f ON p.empl_ssn = f.empl_ssn 
    JOIN random AS r ON p.empl_ssn = r.empl_ssn 
ORDER BY depend_no; 

Answer 1

私はこれ（あなたが発注ではなく、従業員がリストの一番上にあることが必要であるという事実以外のランク（）を探しているようにそれはそう？）のようなものだろう。

を

SELECT 
    empl_ssn 
    , cov_ind_ssn 
    , relationship_code 
    , ROW_NUMBER() OVER (ORDER BY CASE WHEN empl_ssn = cov_ind_ssn THEN 0 ELSE 1 END, birthdate) depend_no 
FROM full_data 
WHERE empl_ssn = @current_employee 

Answer 2

ランク付けアプローチでcase expressionを使用してください。

SELECT 
     empl_ssn 
     , cov_ind_ssn 
     , relationship_code 
     , CASE WHEN empl_ssn = cov_ind_ssn THEN 1 ELSE 
       ROW_NUMBER() OVER (ORDER BY f.birthdate) + 1 
     END AS [depend_no] 
    FROM full_data AS f 
    WHERE empl_ssn = @current_employee