騰訊音樂SQL題

1. 計算歌曲完播率

請根據 user_listen_record、song_library計算出QQ音樂20230306歌曲完播率(播放時長>=聽歌時長)輸出表結構如下,其中完播率保留小數點後2位小數並按照完播率重小到大排序:

song_playback_history

img
已知QQ音樂部分用戶聽歌流水錶格式和樣例數據如下: user_listen_record img 其中ftime爲數據分區時間,uin爲用戶賬號(唯一標識),os_type爲設備端分類,song_id爲歌曲id,app_ver爲應用版本,play_duration爲聽歌時長(秒)

曲庫信息表:song_library img 其中song_id爲歌曲id(唯一標識),song_name歌曲名稱,duration爲歌曲時長(秒),artist_id爲歌手id,artist_name爲歌手名

示例1

輸入例子:

-- ----------------------------
-- Table structure for user_listen_record
-- ----------------------------
DROP TABLE IF EXISTS `user_listen_record`;
CREATE TABLE `user_listen_record` (
  `ftime` bigint(20) DEFAULT NULL,
  `uin` varchar(255) DEFAULT NULL,
  `os_type` varchar(255) DEFAULT NULL,
  `song_id` bigint(20) DEFAULT NULL,
  `app_ver` varchar(255) DEFAULT NULL,
  `play_duration` bigint(20) DEFAULT NULL
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4;

-- ----------------------------
-- Records of user_listen_record
-- ----------------------------
BEGIN;
INSERT INTO `user_listen_record` (`ftime`, `uin`, `os_type`, `song_id`, `app_ver`, `play_duration`) VALUES (20230306, 'a', 'ios', 1001, '10.0.1', 140);
INSERT INTO `user_listen_record` (`ftime`, `uin`, `os_type`, `song_id`, `app_ver`, `play_duration`) VALUES (20230306, 's', 'android', 1001, '10.0.1', 170);
INSERT INTO `user_listen_record` (`ftime`, `uin`, `os_type`, `song_id`, `app_ver`, `play_duration`) VALUES (20230306, 'm', 'ios', 1003, '10.0.5', 100);
INSERT INTO `user_listen_record` (`ftime`, `uin`, `os_type`, `song_id`, `app_ver`, `play_duration`) VALUES (20230306, 'u', 'android', 1004, '10.0.1', 229);
INSERT INTO `user_listen_record` (`ftime`, `uin`, `os_type`, `song_id`, `app_ver`, `play_duration`) VALUES (20230306, 'm', 'ios', 1002, '10.0.5', 230);
INSERT INTO `user_listen_record` (`ftime`, `uin`, `os_type`, `song_id`, `app_ver`, `play_duration`) VALUES (20230306, 'a', 'ios', 1003, '10.0.1', 257);
INSERT INTO `user_listen_record` (`ftime`, `uin`, `os_type`, `song_id`, `app_ver`, `play_duration`) VALUES (20230306, 'u', 'android', 1001, '10.0.1', 290);
INSERT INTO `user_listen_record` (`ftime`, `uin`, `os_type`, `song_id`, `app_ver`, `play_duration`) VALUES (20230306, 's', 'android', 1003, '10.0.1', 170);
INSERT INTO `user_listen_record` (`ftime`, `uin`, `os_type`, `song_id`, `app_ver`, `play_duration`) VALUES (20230306, 'a', 'ios', 1004, '10.0.1', 229);
COMMIT;

DROP TABLE IF EXISTS `song_library`;
CREATE TABLE `song_library` (
  `song_id` bigint(20) DEFAULT NULL,
  `song_name` varchar(255) DEFAULT NULL,
  `duration` bigint(20) DEFAULT NULL,
  `artist_id` bigint(20) DEFAULT NULL,
  `artist_name` varchar(255) DEFAULT NULL
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4;

-- ----------------------------
-- Records of song_library
-- ----------------------------
BEGIN;
INSERT INTO `song_library` (`song_id`, `song_name`, `duration`, `artist_id`, `artist_name`) VALUES (1001, '七里香', 297, 1, '周杰倫');
INSERT INTO `song_library` (`song_id`, `song_name`, `duration`, `artist_id`, `artist_name`) VALUES (1002, '逆戰', 230, 235, '張傑');
INSERT INTO `song_library` (`song_id`, `song_name`, `duration`, `artist_id`, `artist_name`) VALUES (1003, '烏梅子醬', 257, 23, '李榮浩');
INSERT INTO `song_library` (`song_id`, `song_name`, `duration`, `artist_id`, `artist_name`) VALUES (1004, '倒數', 229, 25, '鄧紫棋');
COMMIT;

輸出例子:

ftime|song_id|song_name|play_comp_rate
20230306|1002|逆戰|1.00
20230306|1004|倒數|1.00
20230306|1003|烏梅子醬|0.33
20230306|1001|七里香|0.00

答案

select
    ftime,
    U.song_id as song_id,
    song_name,
    round(
        avg(if (play_duration >= duration, 1, 0)),
        2
    ) as play_comp_rate
from
    user_listen_record U
    join song_library S on U.song_id = S.song_id
where ftime = '20230306'
group by
    ftime,
    U.song_id,
    song_name
order by
    play_comp_rate desc

2. 聽歌時長前3名

請根據 user_listen_record按照每個用戶對歌曲的聽歌時長,排出每個用戶播放前3名歌曲(相同排名取song_id更小的歌曲),最後結果按用戶賬號(uin)從大到小,排名從1到3排序,輸出表結構如下:

user_play_rank

img

輸出例子:

uin|song_id|rank
u|1001|1
u|1004|2
s|1001|1
s|1003|2
m|1002|1
m|1003|2
a|1003|1
a|1004|2
a|1001|3

答案代碼:

SELECT
	uin,
	song_id,
	rank 
FROM
	(
	SELECT
		uin,
		U.song_id,
		row_number() over ( PARTITION BY uin ORDER BY total_time ) AS rank 
	FROM
		(
		SELECT
			uin,
			sum(play_duration) AS total_time,
			U.song_id AS song_id 
		FROM
			user_listen_record U
			JOIN song_library S ON U.song_id = S.song_id 
		GROUP BY
			uin,
			U.song_id 
		) t1 
	) t2 
WHERE
	rank <= 3

3. 每個月Top3的周杰倫歌曲

輸入例子:

drop table if exists play_log;
create table `play_log` (
    `fdate` date,
    `user_id` int,
    `song_id` int
);
insert into play_log(fdate, user_id, song_id)
values 
('2022-01-08', 10000, 0),
('2022-01-16', 10000, 0),
('2022-01-20', 10000, 0),
('2022-01-25', 10000, 0),
('2022-01-02', 10000, 1),
('2022-01-12', 10000, 1),
('2022-01-13', 10000, 1),
('2022-01-14', 10000, 1),
('2022-01-10', 10000, 2),
('2022-01-11', 10000, 3),
('2022-01-16', 10000, 3),
('2022-01-11', 10000, 4),
('2022-01-27', 10000, 4),
('2022-02-05', 10000, 0),
('2022-02-19', 10000, 0),
('2022-02-07', 10000, 1),
('2022-02-27', 10000, 2),
('2022-02-25', 10000, 3),
('2022-02-03', 10000, 4),
('2022-02-16', 10000, 4);

drop table if exists song_info;
create table `song_info` (
    `song_id` int,
    `song_name` varchar(255),
    `singer_name` varchar(255)
);
insert into song_info(song_id, song_name, singer_name) 
values
(0, '明明就', '周杰倫'),
(1, '說好的幸福呢', '周杰倫'),
(2, '江南', '林俊杰'),
(3, '大笨鐘', '周杰倫'),
(4, '黑鍵', '林俊杰');

drop table if exists user_info;
create table `user_info` (
    `user_id`   int,
    `age`       int
);
insert into user_info(user_id, age) 
values
(10000, 18)

輸出例子:

month|ranking|song_name|play_pv
1|1|明明就|4
1|2|說好的幸福呢|4
1|3|大笨鐘|2
2|1|明明就|2
2|2|說好的幸福呢|1
2|3|大笨鐘|1

例子說明:

1月被18-25歲用戶播放次數最高的三首歌爲“明明就”、“說好的幸福呢”、“大笨鐘”,“明明就”和“說好的幸福呢”播放次數相同,排名先後由兩者的song_id先後順序決定。2月同理。

答案:

select
    month,
    ranking,
    song_name,
    play_pv
from
    (
        select
            month,
            row_number() over (
                partition by
                    month
                order by
                    play_pv desc,
                    song_id
            ) as ranking,
            song_name,
            play_pv
        from
            (
                select
                    month (fdate) as month,
                    song_name,
                    PS.song_id as song_id,
                    count(*) as play_pv
                from
                    play_log PS
                    join song_info S on PS.song_id = S.song_id
                    join user_info U on PS.user_id = U.user_id
                where
                    year (fdate) = 2022
                    and age >= 18
                    and age <= 25
                    and singer_name = '周杰倫'
                group by
                    month,
                    song_name,
                    PS.song_id
            ) t1
    ) t2
where
    ranking < 4

4. 語種播放量前三高所有歌曲

表:songplay

+--------------+---------+
| Column Name | Type |
+--------------+---------+
| id | int |
| playcnt | int |
|languageid | int |
+--------------+---------+
id是該表的主鍵列。
languageid是songplay表中ID的外鍵。
該表的每一行都表示歌曲的ID、播放量,語種id。

表: language

+-------------+---------+
| Column Name | Type |
+-------------+---------+
| id | int |
| name | varchar |
+-------------+---------+
Id是該表的主鍵列。
該表的每一行表示語種ID和語種名。

示例1

輸入例子:

drop table if exists  songplay;
create table `songplay`(
`id` int,
`playcnt` int,
`languageid` int
);
insert into songplay
values(1,85001,1);
insert into songplay 
values(2,80001,2);
insert into  songplay 
values(3,60001,2);
insert into  songplay 
values(4,90001,1);
insert into  songplay
values(5,69001,1);
insert into  songplay
values(6,85001,1);
insert into  songplay
values(7,70001,1);
drop table if exists language;
create table `language`(
`id` int,
`name` varchar(255)
);
insert into  language 
values(1,'中文');
insert into  language
values(2,'英文');

輸出例子:

language_name|songid|playcnt
中文|4|90001
中文|1|85001
中文|6|85001
中文|7|70001
英文|2|80001
英文|3|60001

代碼:

開始搞錯思路了,看這個樣例還以爲是求播放總量前三的語種下的所有歌曲,實際上是求每個語種播放量前三的歌曲(存在並列現象)

select
    language_name,
    songid,
    playcnt
from
    (
        select
            songplay.id as songid,
            name as language_name,
            playcnt,
            dense_rank() over (
                PARTITION BY
                    languageid
                ORDER BY
                    playcnt desc
            ) AS rk
        from
            songplay
            join language on songplay.languageid = language.id
    )t1
    where rk <=3

5. 最長連續登錄天數

你正在搭建一個用戶活躍度的畫像,其中一個與活躍度相關的特徵是“最長連續登錄天數”, 請用SQL實現“2023年1月1日-2023年1月31日用戶最長的連續登錄天數”

示例1

輸入例子:

drop table if exists tb_dau;
create table `tb_dau` (
    `fdate` date,
    `user_id` int
);
insert into tb_dau(fdate, user_id)
values 
('2023-01-01', 10000),
('2023-01-02', 10000),
('2023-01-04', 10000);

輸出例子:

user_id|max_consec_days
10000|2

例子說明:

id爲10000的用戶在1月1日及1月2日連續登錄2日,1月4日登錄1日,故最長連續登錄天數爲2日

答案

SELECT
    user_id,
    max(consecutive_day) AS max_consec_days
FROM
    (
        SELECT
            user_id,
            count(diff) AS consecutive_day
        FROM
            (
                SELECT
                    user_id,
                    fdate - rn AS diff
                FROM
                    (
                        SELECT
                            user_id,
                            fdate,
                            row_number() over (PARTITION BY user_id ORDER BY fdate) AS rn
                        FROM
                            tb_dau
                    ) t1
            ) t2
        GROUP BY
            user_id,
            diff
    ) t3
GROUP BY
    user_id

6.SQL實現文本處理

現有試卷信息表examination_info(exam_id試卷ID, tag試卷類別, difficulty試卷難度, duration考試時長):
+----+---------+------------------+------------+----------+---------------------+
| id | exam_id | tag | difficulty | duration |
+----+---------+------------------+------------+----------+---------------------+
| 1 | 9001 | 算法 | hard | 60 |
| 2 | 9002 | 算法 | hard | 80 |
| 3 | 9003 | SQL | medium | 70 |
| 4 | 9004 | 算法,medium,80 | | 0 |
+----+---------+------------------+------------+----------+---------------------+

錄題同學有一次手誤將部分記錄的試題類別tag、難度、時長同時錄入到了tag字段,
請幫忙找出這些錄錯了的記錄,並拆分後按正確的列類型輸出。
由示例數據結果輸出如下:
+---------+--------+------------+----------+
| exam_id | tag | difficulty | duration |
+---------+--------+------------+----------+
| 9004 | 算法 | medium | 80 |
+---------+--------+------------+----------+

示例1

輸入例子:

drop table if exists examination_info,exam_record;
CREATE TABLE examination_info (
    id int PRIMARY KEY AUTO_INCREMENT COMMENT '自增ID',
    exam_id int UNIQUE NOT NULL COMMENT '試卷ID',
    tag varchar(32) COMMENT '類別標籤',
    difficulty varchar(8) COMMENT '難度',
    duration int NOT NULL COMMENT '時長',
    release_time datetime COMMENT '發佈時間'
)CHARACTER SET utf8 COLLATE utf8_general_ci;

INSERT INTO examination_info(exam_id,tag,difficulty,duration,release_time) VALUES
  (9001, '算法', 'hard', 60, '2020-01-01 10:00:00'),
  (9002, '算法', 'hard', 80, '2020-01-01 10:00:00'),
  (9003, 'SQL', 'medium', 70, '2020-01-01 10:00:00'),
  (9004, '算法,medium,80','', 0, '2020-01-01 10:00:00');

輸出例子:

exam_id|tag|difficulty|duration
9004|算法|medium|80

答案代碼

牛客原題:https://www.nowcoder.com/practice/a5475ed3b5ab4de58e2ea426b4b2db76

SELECT
    exam_id,
    -- 查找字段tag中','這個字符的每一個位置並排序,截取第一個','向左所有的字符。 
    substring_index (tag, ',', 1) as tag,
    -- difficult在中間位置,需要截取2次
    substring_index (substring_index (tag, ',', 2), ',', -1) as difficult,
    -- 查找字段tag中','這個字符的每一個位置並排序,截取最後1個','向右所有的字符。並且轉換數據格式。  
    substring_index (tag, ',', -1) as duration
from
    examination_info
where
    -- 定位到出現串列的數據
    tag like '%,%'
發表評論
所有評論
還沒有人評論,想成為第一個評論的人麼? 請在上方評論欄輸入並且點擊發布.
相關文章