SQL Server面试题

来自牛奶河Wiki
跳到导航 跳到搜索

统计销售最差的车

车辆按批销售，每次销售若干辆同型号的车，表中就记录增加一条记录。

问：查询总销售量大于100，且总销售辆最少的3个型号的车及其总销售量。

MODEL CNT
A 20
B 50
B 100
C 900
C 500
D 400
E 200
F 40
G 300 
/* Q1: Temp Table 车俩销售日志
 * 若含有中文数值，改为 nvarchar
 *
 * 2023/8/8 Adam
 */
create table #tmp_salelog (
   d_model   varchar(8),
   v_cnt     int
)
;

insert into #tmp_salelog values
('A', 20),
('B', 50),
('B', 100),
('C', 900),
('C', 500),
('D', 400),
('E', 200),
('F', 40),
('G', 300)
;

-- 问：查询总销售量大于100，且总销售辆最少的3个型号的车及其总销售量
select   top 3
         d_model, sum(v_cnt) v_cnt_total
from     #tmp_salelog
group by d_model
having   sum(v_cnt) > 100
order by v_cnt_total
;

游戏打包销售折扣

销售平台进行游戏打包促销。将任意个游戏打包为一组，根据游戏数量制定折扣。打包的游戏数量限定2个至4个。当包含2个游戏时折扣为9折，3个时8折，4个时6折。

问1：计算有多少种(个数)不同的打包组合方式

问2：打包购买游戏时，分别计算出2个打包，3个打包和4个打包时价格最贵的包。要求最终用一个查询得出结果 

/* Q2: Temp Table 游戏价格
 * 若含有中文数值，改为 nvarchar
 * 2023/8/8 Adam
 */
create table #tmp_games (
   d_gname   varchar(8),
   v_amt     numeric(8,2)
)
;

insert into #tmp_games values
('A', 29.33),
('B', 19.22),
('C', 25.81),
('D', 16.79),
('E', 20.78),
('F', 25.32)
;

/*
 * tmp_games 中数据，关于 gname 应该唯一。否则需要预警及去重
 */
-- 问1：计算有多少种(个数)不同的打包组合方式
-- 如果游戏比较多（大于 300 时，本写法产生的笛卡尔积将达到三千万），需要另外一种的写法：通过排列组合计算

-- 答案一：允许打包重复游戏
select   '2' type_id, count(*) num
from     #tmp_games t1, #tmp_games t2
where    t1.d_gname <= t2.d_gname
union all
select   '3' type_id, count(*) num
from     #tmp_games t1, #tmp_games t2, #tmp_games t3
where    t1.d_gname <= t2.d_gname
and      t2.d_gname <= t3.d_gname
union all
select   '4' type_id, count(*) num
from     #tmp_games t1, #tmp_games t2, #tmp_games t3, #tmp_games t4
where    t1.d_gname <= t2.d_gname
and      t2.d_gname <= t3.d_gname
and      t3.d_gname <= t4.d_gname
;

-- 答案二：不允许打包重复游戏
select   '2' type_id, count(*) num
from     #tmp_games t1, #tmp_games t2
where    t1.d_gname < t2.d_gname
union all
select   '3' type_id, count(*) num
from     #tmp_games t1, #tmp_games t2, #tmp_games t3
where    t1.d_gname < t2.d_gname
and      t2.d_gname < t3.d_gname
union all
select   '4' type_id, count(*) num
from     #tmp_games t1, #tmp_games t2, #tmp_games t3, #tmp_games t4
where    t1.d_gname < t2.d_gname
and      t2.d_gname < t3.d_gname
and      t3.d_gname < t4.d_gname
;


-- 问2：打包购买游戏时，分别计算出2个打包，3个打包和4个打包时价格最贵的包。要求最终用一个查询得出结果
-- 答案一：允许打包重复游戏
select   *
from     (
select   top 1
         '2' type_id,
         t1.d_gname + t2.d_gname d_game_group,
         Convert(decimal(18,2), (t1.v_amt + t2.v_amt)*0.9) amt_total
from     #tmp_games t1, #tmp_games t2
where    t1.d_gname <= t2.d_gname
order by 3 desc
)a
union all
select   *
from     (
select   top 1
         '3' type_id,
         t1.d_gname + t2.d_gname + t3.d_gname d_game_group,
         Convert(decimal(18,2), (t1.v_amt + t2.v_amt + t3.v_amt)*0.8) amt_total
from     #tmp_games t1, #tmp_games t2, #tmp_games t3
where    t1.d_gname <= t2.d_gname
and      t2.d_gname <= t3.d_gname
order by 3 desc
)b
union all
select   *
from     (
select   top 1
         '4' type_id,
         t1.d_gname + t2.d_gname + t3.d_gname + t4.d_gname d_game_group,
         Convert(decimal(18,2), (t1.v_amt + t2.v_amt + t3.v_amt + t4.v_amt)*0.6) amt_total
from     #tmp_games t1, #tmp_games t2, #tmp_games t3, #tmp_games t4
where    t1.d_gname <= t2.d_gname
and      t2.d_gname <= t3.d_gname
and      t3.d_gname <= t4.d_gname
order by 3 desc
)c




-- 答案二：不允许打包重复游戏
select   *
from     (
select   top 1
         '2' type_id,
         t1.d_gname + t2.d_gname d_game_group,
         Convert(decimal(18,2), (t1.v_amt + t2.v_amt)*0.9) amt_total
from     #tmp_games t1, #tmp_games t2
where    t1.d_gname < t2.d_gname
order by 3 desc
)a
union all
select   *
from     (
select   top 1
         '3' type_id,
         t1.d_gname + t2.d_gname + t3.d_gname d_game_group,
         Convert(decimal(18,2), (t1.v_amt + t2.v_amt + t3.v_amt)*0.8) amt_total
from     #tmp_games t1, #tmp_games t2, #tmp_games t3
where    t1.d_gname < t2.d_gname
and      t2.d_gname < t3.d_gname
order by 3 desc
)b
union all
select   *
from     (
select   top 1
         '4' type_id,
         t1.d_gname + t2.d_gname + t3.d_gname + t4.d_gname d_game_group,
         Convert(decimal(18,2), (t1.v_amt + t2.v_amt + t3.v_amt + t4.v_amt)*0.6) amt_total
from     #tmp_games t1, #tmp_games t2, #tmp_games t3, #tmp_games t4
where    t1.d_gname < t2.d_gname
and      t2.d_gname < t3.d_gname
and      t3.d_gname < t4.d_gname
order by 3 desc
)c

经销商代理品牌额度统计

DEALER_INFO 表在每个经销商申请代理一个品牌的额度时，单独维护对应的基本信息

CREDIT_LIMIT_INFO 表维护每个授信协议的额度金额

问：要求以相同的证件号码为唯一标识识别为一个客户，取每个客户最近一次维护的基本信息进行报送，同时取这个客户已用额度总和。

DEALER_INFO
经销商授信协议号码 经销商名称 经销商证件号 注册地址 员工人数 信息维护日期
DEALER_NUMBER DEALER_NAME DEALER_ID_NO ADDRESS STUFF_NUMBER MODIFY_DATE
1001001 TEST_01 ID_001 BEIJING CHAOYANG 100 2020-01-01
2001001 TEST_01 ID_001 BEIJING CHAOYANG 200 2020-02-01
1002002 TEST_02 ID_002 SHANGHAI PUDONG 1000 2020-03-15
3002002 TEST_02 ID_002 SHANGHAI MINHANG 1000 2020-01-20
2003003 TEST_03 ID_003 BEIJING HAIDIAN 50 2020-02-25
4003003 TEST_03 ID_003 BEIJING DONGCHENG 200 2020-01-01
1004004 TEST_04 ID_004 SHANGHAI HUANGPU 100 2020-03-01
CREDIT_LIMIT_INFO
经销商授信协议号码 品牌 已用额度
DEALER_NUMBER BRAND UTILIZED_LIMIT
1001001 FAW 1000
2001001 SKODA 1500
1002002 FAW 200
3002002 AUDI 400
2003003 SVW 500
4003003 PORCHE 50
1004004 FAW 700

最终报文数据样式

客户名称 证件号码 注册地址 员工人数 已用额度
DEALER_NAME DEALER_ID_NO ADDRESS STUFF_NUMBER UTILIZED_LIMIT
TEST_01 ID_001 BEIJING CHAOYANG 200 2500
TEST_02 ID_002 SHANGHAI PUDONG 1000 600
TEST_03 ID_003 BEIJING HAIDIAN 50 550
TEST_04 ID_004 SHANGHAI HUANGPU 100 700 
/* Q3: Temp Table 经销商信息
 * 若含有中文数值，改为 nvarchar
 * 2023/8/8 Adam
 */
create table #tmp_dealer_info (
   d_dealer_number    varchar(8),
   d_dealer_name      varchar(100),
   d_dealer_id_no     varchar(8),
   d_address          varchar(200),
   v_stuff_number     INT,
   d_modify_date      date
)
;

insert into #tmp_dealer_info
values
('1001001','TEST_01','ID_001','BEIJING CHAOYANG',100,'2020-01-01'),
('2001001','TEST_01','ID_001','BEIJING CHAOYANG',200,'2020-02-01'),
('1002002','TEST_02','ID_002','SHANGHAI PUDONG',1000,'2020-03-15'),
('3002002','TEST_02','ID_002','SHANGHAI MINHANG',1000,'2020-01-20'),
('2003003','TEST_03','ID_003','BEIJING HAIDIAN',50,'2020-02-25'),
('4003003','TEST_03','ID_003','BEIJING DONGCHENG',200,'2020-01-01'),
('1004004','TEST_04','ID_004','SHANGHAI HUANGPU',100,'2020-03-01')
;

create table #tmp_credit_limit_info (
   d_dealer_number    varchar(8),
   d_brand            varchar(100),
   v_utilized_limit   INT
)
;

insert into #tmp_credit_limit_info
values
('1001001','FAW',1000),
('2001001','SKODA',1500),
('1002002','FAW',200),
('3002002','AUDI',400),
('2003003','SVW',500),
('4003003','PORCHE',50),
('1004004','FAW',700)

-- 问：要求以相同的证件号码为唯一标识识别为一个客户，取每个客户最近一次维护的基本信息进行报送，同时取这个客户已用额度总和。
-- 证件号码为唯一标识，为两表关联条件、分组排序条件
-- P.S. 有可能出现""协议号码""未使用额度（无记录），显示为零
select   --
         o1.d_dealer_name,
         o1.d_dealer_id_no,
         o1.d_address,
         o1.v_stuff_number,
         o2.v_utilized_limit_total
from     (
         -- 按经销商证件号分组，倒排信息维护日期, rn=1 为最后修改记录
         select   d_dealer_number,
                  d_dealer_name,
                  d_dealer_id_no,
                  d_address,
                  v_stuff_number,
                  d_modify_date,
                  row_number() over(partition BY d_dealer_id_no order BY d_modify_date desc) rn
         from     #tmp_dealer_info
         ) o1,
        (
        -- 客户已用额度总和，若未使用额度（无记录），显示为零
        select   t1.d_dealer_id_no,
                 isnull(sum(t2.v_utilized_limit), 0)  v_utilized_limit_total
        from     #tmp_dealer_info t1
        left join #tmp_credit_limit_info t2 on t1.d_dealer_number = t2.d_dealer_number
        group by t1.d_dealer_id_no
        ) o2
where    o1.d_dealer_id_no = o2.d_dealer_id_no
and      o1.rn = 1
;
取自“https://mwbbs.eu.org/wiki/index.php?title=SQL_Server面试题&oldid=1539”