我搜索了应用程序，找到了从itemregistrationpangkat表中按某些标准（类别，经验等）过滤的个人信息：
itemregistrationpangkatID|itemregistrationID|NegeriID|yeartamatkhidmat|yearmulakhidmat
--------------------------------------------------------------------------
1 | 2 | Kedah | 2001 | 2002
2 | 2 | Kedah | 2003 | 2007
3 | 3 | Melaka | 2008 | 2012
4 | 3 | Melaka | 2013 | 2018
5 | 4 | KL | 2000 | 2001
itemregistrationID | name | SectionID | CategoryID
-----------------------------------------------------------
1 | Eric | 1 | 2
2 | Tom | 2 | 4
3 | Anne | 2 | 4
4 | Marie | 3 | 2
5 | Bill | 1 | 3
我对最后一个过滤器“ negeri的tempoh perkhidmatan”有疑问。我需要按州（negeri）计算工作经验的数量。例如，当搜索州（negeri）“ x”中的5年人员时，sql将汇总所选州每个人的经验年限。
这是按条件搜索SQL的完整代码：
$query = DB::table('itemregistrations')
->join('sections', 'itemregistrations.SectionID', '=', 'sections.SectionID')
->join('categories', 'itemregistrations.CategoryID', '=', 'categories.CategoryID')
->join('operasi', 'itemregistrations.OperasiID', '=', 'operasi.OperasiID')
->join('negeri', 'itemregistrations.NegeriID', '=', 'negeri.NegeriID')
->join('gred', 'itemregistrations.GredID', '=', 'gred.GredID')
->where('itemregistrations.statusProID', '=', 1)
->select('itemregistrations.name','sections.sectionname', 'categories.categoryname', 'operasi.operasiname', 'itemregistrations.Nobadan', 'itemregistrations.lahir_yy', 'itemregistrations.pdrm_yy', 'gred.namagred', 'itemregistrations.itemRegistrationID', '');
if($request->input('negeri_lahir') != ''){
$query->where('itemregistrations.NegeriID', $request->input('negeri_lahir'));
}
if($request->input('kategori') != '') {
$query->where('itemregistrations.CategoryID', $request->input('kategori'));
}
if($request->input('pangkat') != '') {
$query->where('itemregistrations.OperasiID', $request->input('pangkat'));
}
if(request('umur')) {
$query->whereRaw('YEAR(CURDATE()) - lahir_yy >= ?', [request('umur')]);
}
if($request->input('gred') != '') {
$query->where('itemregistrations.GredID', $request->input('gred'));
}
if(request('tempoh')) {
$query->whereRaw('YEAR(CURDATE()) - pdrm_yy >= ?', [request('tempoh')]);
}
if($request->input('negeri_perkhidmatan') != '') {
$query->join('itemregistrationpangkat', 'itemregistrationpangkat.itemRegistrationID', '=', 'itemregistrations.itemRegistrationID')
->where('itemregistrationpangkat.NegeriID', $request->input('negeri_perkhidmatan'));
}
if(request('tempoh_negeri')) {
$query->select(DB::raw('m.itemRegistrationID, sum(m.duration)'))
->from(DB::raw('(SELECT itemRegistrationID, NegeriID, yeartamatkhidmat - yearmulakhidmat as duration FROM itemregistrationpangkat) AS m
RIGHT JOIN itemregistrations ON itemregistrations.itemRegistrationID=m.itemRegistrationID'))
->distinct()
->groupBy('m.itemRegistrationID');
}
$newitem = $query->get();
return response::json($newitem);
涉及解决的代码是这个（最后一个过滤器）：
if(request('tempoh_negeri')) {
$query->select(DB::raw('m.itemRegistrationID, sum(m.duration)'))
->from(DB::raw('(SELECT itemRegistrationID, NegeriID, yeartamatkhidmat - yearmulakhidmat as duration FROM itemregistrationpangkat) AS m
RIGHT JOIN itemregistrations ON itemregistrations.itemRegistrationID=m.itemRegistrationID'))
->distinct()
->groupBy('m.itemRegistrationID');
}
问题是重复的行。 上面的控制台日志显示：
0:
itemRegistrationID: 791
sum(m.duration): 6
应该是：
0:
itemRegistrationID: 791
sum(m.duration): 3
distinct（）无法与groupby一起使用，如果我删除groupby，它可以删除重复的行。
答案 0 :(得分：1)
我认为它显示重复的原因是因为RIGHT JOIN。我本人已经测试了相同的条件，下面是我的测试示例：
--Lets say my itemregistrations table look like this:
+----------------------+-------------+------------+
| itemRegistrationID | firstname | lastname |
+----------------------+-------------+------------+
| 1 | A | B |
| 2 | C | D |
| 3 | E | F |
| 1 | A | A |
+----------------------+-------------+------------+
-- notice that there are repeated value of = 1 in itemregistrationID
-- And my itemregistrationpangkat table look like this:
+----------------------+------------+--------------------+-------------------+
| itemRegistrationID | NegeriID | yeartamatkhidmat | yearmulakhidmat |
+----------------------+------------+--------------------+-------------------+
| 1 | J | 2019-04-26 | 2009-12-01 |
| 2 | S | 2018-12-31 | 1999-08-19 |
| 3 | P | 2017-12-31 | 2001-03-02 |
+----------------------+------------+--------------------+-------------------+
我在上面提取了您的MySQL查询，使其变成了这样：
SELECT m.itemRegistrationID, SUM(m.duration)
FROM
(SELECT itemRegistrationID, NegeriID,
YEAR(yeartamatkhidmat) - YEAR(yearmulakhidmat) AS duration
FROM itemregistrationpangkat) AS m
RIGHT JOIN itemregistrations
ON itemregistrations.itemRegistrationID=m.itemRegistrationID
GROUP BY m.itemRegistrationID;
-- with the query above, my results are as following:
+----------------------+-------------------+
| itemRegistrationID | sum(m.duration) |
+----------------------+-------------------+
| 1 | 20 |
| 2 | 19 |
| 3 | 16 |
+----------------------+-------------------+
为了查看导致id = 1重复值的原因，我进行如下查询：
SELECT * FROM
(SELECT itemRegistrationID, NegeriID,
YEAR(yeartamatkhidmat) - YEAR(yearmulakhidmat) AS duration
FROM itemregistrationpangkat) AS m
RIGHT JOIN itemregistrations ON
itemregistrations.itemRegistrationID=m.itemRegistrationID;
-- and it shows the following result:
+----------------------+------------+------------+----------------------+-------------+------------+
| itemRegistrationID | NegeriID | duration | itemRegistrationID | firstname | lastname |
+----------------------+------------+------------+----------------------+-------------+------------+
| 1 | J | 10 | 1 | A | B |
| 1 | J | 10 | 1 | A | A |
| 2 | S | 19 | 2 | C | D |
| 3 | P | 16 | 3 | E | F |
+----------------------+------------+------------+----------------------+-------------+------------+
当然，在您的SUM(DISTINCT m.duration)中添加SELECT会删除所有m.duration重复项。这就是我担心的地方。我不认为m.duration应该是决定重复项的因素，因为您有itemregistrationID和NegeriID，因此两者的结合应该是唯一的（或者至少是其他领域的结合，这两个将是唯一的）。我进一步的测试表明，如果itemregistrationpangkat表中有两个itemRegistrationID等不同的NegeriID，则SUM(DISTINCT m.duration)将不起作用。如果我找到更好的方法，让我尝试一些事情并更新我的答案。
更新： 我在这里创建了一个小提琴：DB Fiddle