Skip to content

Commit b8fa7de

Browse files
myoshi2891myoshi2891
authored
Merge pull request #186 from myoshi2891/dev/macbook_pro
SQL: Basic select 586. Customer Placing the Largest Number of Orders
2 parents 2a04b56 + d9194ef commit b8fa7de

3 files changed

Lines changed: 611 additions & 0 deletions

File tree

‎SQL/Leetcode/Basic select/586. Customer Placing the Largest Number of Orders/gpt/Customer_Placing_the_Largest_Number_of_Orders_mysql.ipynb‎

Lines changed: 227 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,227 @@
1+
{
2+
"cells": [
3+
{
4+
"cell_type": "markdown",
5+
"id": "eeeadbe0",
6+
"metadata": {},
7+
"source": [
8+
"# MySQL 8.0.40\n",
9+
"\n",
10+
"## 0) 前提\n",
11+
"\n",
12+
"* エンジン: **MySQL 8**\n",
13+
"* 並び順: 任意（`ORDER BY` を付けない）\n",
14+
"* `NOT IN` は NULL 罠のため回避\n",
15+
"* 判定は **ID 基準**、表示は仕様どおりの列名と順序\n",
16+
"\n",
17+
"## 1) 問題\n",
18+
"\n",
19+
"* `Orders` から **最も多く注文を行った顧客の `customer_number`** を返す\n",
20+
" （テストでは最大が一意。Follow up: 同数最大が複数いても全件返す）\n",
21+
"\n",
22+
"* 入力テーブル例:\n",
23+
"\n",
24+
" ```markdown\n",
25+
" Table: Orders\n",
26+
" +-----------------+----------+\n",
27+
" | Column Name | Type |\n",
28+
" +-----------------+----------+\n",
29+
" | order_number | int | -- PK\n",
30+
" | customer_number | int |\n",
31+
" +-----------------+----------+\n",
32+
" ```\n",
33+
"\n",
34+
"* 出力仕様:\n",
35+
"\n",
36+
" ```markdown\n",
37+
" +-----------------+\n",
38+
" | customer_number |\n",
39+
" +-----------------+\n",
40+
" ```\n",
41+
"\n",
42+
"## 2) 最適解（単一クエリ）\n",
43+
"\n",
44+
"> **ウィンドウ関数＋事前集計**で 1 クエリ。`OVER` 内の `ORDER BY` は順位付けのためで、最終 `SELECT` に `ORDER BY` は不要。\n",
45+
"\n",
46+
"```sql\n",
47+
"WITH cnt AS (\n",
48+
" SELECT\n",
49+
" customer_number,\n",
50+
" COUNT(*) AS order_cnt\n",
51+
" FROM Orders\n",
52+
" GROUP BY customer_number\n",
53+
"),\n",
54+
"win AS (\n",
55+
" SELECT\n",
56+
" customer_number,\n",
57+
" DENSE_RANK() OVER (ORDER BY order_cnt DESC) AS rnk\n",
58+
" FROM cnt\n",
59+
")\n",
60+
"SELECT\n",
61+
" customer_number\n",
62+
"FROM win\n",
63+
"WHERE rnk = 1;\n",
64+
"\n",
65+
"Runtime 429 ms\n",
66+
"Beats 75.41%\n",
67+
"\n",
68+
"```\n",
69+
"\n",
70+
"* これで **一意最大**も**同数最大が複数**も対応（Follow up 充足）\n",
71+
"\n",
72+
"## 3) 代替解\n",
73+
"\n",
74+
"> **最大値をサブクエリで求めて一致フィルタ**。ウィンドウが重い環境や互換用に。\n",
75+
"\n",
76+
"```sql\n",
77+
"WITH cnt AS (\n",
78+
" SELECT customer_number, COUNT(*) AS order_cnt\n",
79+
" FROM Orders\n",
80+
" GROUP BY customer_number\n",
81+
"),\n",
82+
"mx AS (\n",
83+
" SELECT MAX(order_cnt) AS max_cnt FROM cnt\n",
84+
")\n",
85+
"SELECT c.customer_number\n",
86+
"FROM cnt AS c\n",
87+
"JOIN mx ON c.order_cnt = mx.max_cnt;\n",
88+
"\n",
89+
"Runtime 469 ms\n",
90+
"Beats 41.83%\n",
91+
"\n",
92+
"```\n",
93+
"\n",
94+
"※ `NOT IN` は未使用。`ORDER BY ... LIMIT 1` でも実現できるが、本要件では**結果順は任意**かつ **最大同率全件**を自然に返せる上記方式が安全。\n",
95+
"\n",
96+
"## 4) 要点解説\n",
97+
"\n",
98+
"* **方針**: まず `customer_number` 単位で件数を集計 → 上位判定（`DENSE_RANK` か `MAX` 照合） → 仕様列のみ投影。\n",
99+
"* **NULL / 重複**:\n",
100+
"\n",
101+
" * `customer_number` が NULL の行が存在するなら集計前に `WHERE customer_number IS NOT NULL` を入れる（問題文では想定外だが堅牢性の観点）。\n",
102+
" * `order_number` は PK のため重複はなし。\n",
103+
"* **安定性**: 出力順は問わないため **最終 `ORDER BY` は不要**。上位選別はウィンドウ内で完結。\n",
104+
"\n",
105+
"## 5) 計算量（概算）\n",
106+
"\n",
107+
"* `GROUP BY` 集計: **O(N)**～**O(N log N)**（ヒープ/ソート次第）\n",
108+
"* ウィンドウ `DENSE_RANK`（代替は `MAX` 照合）: 集計後のユニーク顧客数を M として **O(M log M)**（内部ソート含む）\n",
109+
" ※ 代替解（`MAX` 照合）は **O(M)** でやや軽量。\n",
110+
"\n",
111+
"## 6) 図解（Mermaid 超保守版）\n",
112+
"\n",
113+
"```mermaid\n",
114+
"flowchart TD\n",
115+
" A[Orders] --> B[customer_number ごとに COUNT]\n",
116+
" B --> C[順位付け DENSE_RANK または MAX 照合]\n",
117+
" C --> D[rnk=1 または order_cnt=max_cnt を抽出]\n",
118+
" D --> E[出力 customer_number]\n",
119+
"```\n"
120+
]
121+
},
122+
{
123+
"cell_type": "markdown",
124+
"id": "4392213e",
125+
"metadata": {},
126+
"source": [
127+
"## 結論（用途別ベスト）\n",
128+
"\n",
129+
"### 1) 最大が一意（本問題の前提）なら最短クエリが最速になりやすい\n",
130+
"\n",
131+
"```sql\n",
132+
"SELECT customer_number\n",
133+
"FROM Orders\n",
134+
"GROUP BY customer_number\n",
135+
"ORDER BY COUNT(*) DESC\n",
136+
"LIMIT 1;\n",
137+
"\n",
138+
"Runtime 470 ms\n",
139+
"Beats 41.18%\n",
140+
"\n",
141+
"```\n",
142+
"\n",
143+
"* 余計な CTE・結合・ウィンドウ不要。\n",
144+
"* `GROUP BY` 後に件数降順で **先頭 1 件だけ**返すので、実装や実行計画的にもシンプル。\n",
145+
"\n",
146+
"### 2) 同数最大をすべて返したい（Follow-up 汎用）\n",
147+
"\n",
148+
"**ウィンドウ関数なし**で 1 クエリ：\n",
149+
"\n",
150+
"```sql\n",
151+
"SELECT customer_number\n",
152+
"FROM Orders\n",
153+
"GROUP BY customer_number\n",
154+
"HAVING COUNT(*) = (\n",
155+
" SELECT COUNT(*) AS mx\n",
156+
" FROM Orders\n",
157+
" GROUP BY customer_number\n",
158+
" ORDER BY mx DESC\n",
159+
" LIMIT 1\n",
160+
");\n",
161+
"\n",
162+
"Runtime 433 ms\n",
163+
"Beats 72.43%\n",
164+
"\n",
165+
"```\n",
166+
"\n",
167+
"* 内側サブクエリで「最大件数」だけを 1 行取得 → 外側で一致フィルタ。\n",
168+
"* あなたの `cnt→mx→JOIN` 版よりも結合が無いぶん軽くなることが多いです。\n",
169+
"\n",
170+
"### 3) ウィンドウ関数派（可読性重視）なら CTE をやめて 1 段で\n",
171+
"\n",
172+
"```sql\n",
173+
"SELECT customer_number\n",
174+
"FROM (\n",
175+
" SELECT\n",
176+
" customer_number,\n",
177+
" DENSE_RANK() OVER (ORDER BY COUNT(*) DESC) AS rnk\n",
178+
" FROM Orders\n",
179+
" GROUP BY customer_number\n",
180+
") t\n",
181+
"WHERE rnk = 1;\n",
182+
"\n",
183+
"Runtime 461 ms\n",
184+
"Beats 48.24%\n",
185+
"\n",
186+
"```\n",
187+
"\n",
188+
"* `COUNT(*)` を直接 `ORDER BY` に使い、そのまま `DENSE_RANK`。\n",
189+
"* 中間 `cnt` CTE のマテリアライズを避けられる分だけ有利になる場合があります。\n",
190+
"\n",
191+
"---\n",
192+
"\n",
193+
"## 実務チューニングのヒント\n",
194+
"\n",
195+
"1. **インデックス**\n",
196+
"\n",
197+
"```sql\n",
198+
"CREATE INDEX idx_orders_customer ON Orders(customer_number);\n",
199+
"```\n",
200+
"\n",
201+
"* `GROUP BY customer_number` の集約が大幅にラクになります（全表スキャン回避/ソート削減）。\n",
202+
"\n",
203+
"2. **CTE は必要最小限に**\n",
204+
" MySQL 8 では CTE がマテリアライズされるケースがあり、単回参照の中間表は**派生表**に畳んだ方が速いことが多いです（上の「3）」がそれ）。\n",
205+
"\n",
206+
"3. **LeetCode の Runtime はノイズ大**\n",
207+
" 実環境では **EXPLAIN** で実行計画を確認し、`rows` 見積もり・ファイルソート有無・テンポラリ使用などをチェックしてく さい。\n",
208+
"\n",
209+
"---\n",
210+
"\n",
211+
"## まとめ\n",
212+
"\n",
213+
"* **一意最大前提**なら ⇒ `GROUP BY ... ORDER BY COUNT(*) DESC LIMIT 1` が最有力。\n",
214+
"* **同率最大も返す**なら ⇒ `HAVING COUNT(*) = (SELECT ... LIMIT 1)` がシンプル＆速いことが多い。\n",
215+
"* **ウィンドウ採用**なら ⇒ 中間 CTE 省略の 1 段構成に。\n",
216+
"* **物理対策** ⇒ `customer_number` にインデックス。\n"
217+
]
218+
}
219+
],
220+
"metadata": {
221+
"language_info": {
222+
"name": "python"
223+
}
224+
},
225+
"nbformat": 4,
226+
"nbformat_minor": 5
227+
}

0 commit comments

Comments
 (0)