我在 SQL 中有这个表：
CREATE TABLE myt (
name VARCHAR(50) NOT NULL,
year INTEGER NOT NULL,
gender CHAR(1) NOT NULL CHECK (gender IN ('M', 'F')),
country_of_birth VARCHAR(50) NOT NULL,
age INTEGER NOT NULL CHECK (age > 0),
number_of_football_games_attended INTEGER CHECK (number_of_football_games_attended >= 0),
PRIMARY KEY (name, year)
);
INSERT INTO myt (name, year, gender, country_of_birth, age, number_of_football_games_attended)
VALUES
('John', 2010, 'M', 'USA', 25, 3),
('John', 2015, 'M', 'USA', 30, 8),
('John', 2020, 'M', 'USA', 35, 12),
('Maria', 2012, 'F', 'Brazil', 22, 2),
('Maria', 2014, 'F', 'Brazil', 24, 5),
('Maria', 2019, 'F', 'Brazil', 29, 15),
('Sofia', 2018, 'F', 'Spain', 31, 7),
('Sofia', 2019, 'F', 'Spain', 32, 9),
('Sofia', 2021, 'F', 'Spain', 34, 13)
);
name year gender country_of_birth age number_of_football_games_attended
John 2010 M USA 25 3
John 2015 M USA 30 8
John 2020 M USA 35 12
Maria 2012 F Brazil 22 2
Maria 2014 F Brazil 24 5
Maria 2019 F Brazil 29 15
Sofia 2018 F Spain 31 7
Sofia 2019 F Spain 32 9
Sofia 2021 F Spain 34 13
我的目标是在每个人的最短年份和最长年份之间添加缺失的数据行，并用逻辑值回填列（例如，性别始终保持不变，出生国家始终保持不变，年龄增加 1，参加的足球比赛不适用，因为我们无法从逻辑上推断这一点）
name year gender country_of_birth age number_of_football_games_attended data_source
John 2010 M USA 25 3 IN ORIGINAL DATA
John 2011 M USA 26 MISSING BACKFILLED
John 2012 M USA 27 MISSING BACKFILLED
John 2013 M USA 28 MISSING BACKFILLED
John 2014 M USA 29 MISSING BACKFILLED
John 2015 M USA 30 8 IN ORIGINAL DATA
John 2016 M USA 31 MISSING BACKFILLED
John 2017 M USA 32 MISSING BACKFILLED
John 2018 M USA 33 MISSING BACKFILLED
John 2019 M USA 34 MISSING BACKFILLED
John 2020 M USA 35 12 IN ORIGINAL DATA
Maria 2012 F Brazil 22 2 IN ORIGINAL DATA
Maria 2013 F Brazil 23 MISSING BACKFILLED
Maria 2014 F Brazil 24 5 IN ORIGINAL DATA
Maria 2015 F Brazil 25 MISSING BACKFILLED
Maria 2016 F Brazil 26 MISSING BACKFILLED
Maria 2017 F Brazil 27 MISSING BACKFILLED
Maria 2018 F Brazil 28 MISSING BACKFILLED
Maria 2019 F Brazil 29 15 IN ORIGINAL DATA
Sofia 2018 F Spain 31 7 IN ORIGINAL DATA
Sofia 2019 F Spain 32 9 IN ORIGINAL DATA
Sofia 2020 F Spain 33 MISSING BACKFILLED
Sofia 2021 F Spain 34 13 IN ORIGINAL DATA
我尝试使用间隙和孤岛方法为此编写 SQL 代码（我在 1=1 上使用联接而不是交叉联接，以便这可以在不支持交叉联接的服务器上工作）：
WITH calendar AS (
SELECT 2010 as year UNION ALL
SELECT 2011 UNION ALL
SELECT 2012 UNION ALL
SELECT 2013 UNION ALL
SELECT 2014 UNION ALL
SELECT 2015 UNION ALL
SELECT 2016 UNION ALL
SELECT 2017 UNION ALL
SELECT 2018 UNION ALL
SELECT 2019 UNION ALL
SELECT 2020 UNION ALL
SELECT 2021
),
person_info AS (
SELECT DISTINCT
name,
FIRST_VALUE(gender) OVER (PARTITION BY name ORDER BY year) as gender,
FIRST_VALUE(country_of_birth) OVER (PARTITION BY name ORDER BY year) as country_of_birth,
MIN(year) as min_year,
MAX(year) as max_year,
MIN(age) as start_age
FROM myt
GROUP BY name
),
filled_gaps AS (
SELECT
p.name,
c.year,
p.gender,
p.country_of_birth,
CASE
WHEN m.age IS NOT NULL THEN m.age
ELSE p.start_age + (c.year - p.min_year)
END as age,
COALESCE(
CAST(m.number_of_football_games_attended AS VARCHAR),
'MISSING'
) as number_of_football_games_attended,
CASE
WHEN m.name IS NOT NULL THEN 'IN ORIGINAL DATA'
ELSE 'BACKFILLED'
END as data_source
FROM person_info p
JOIN calendar c ON 1=1
LEFT JOIN myt m
ON m.name = p.name
AND m.year = c.year
WHERE c.year BETWEEN p.min_year AND p.max_year
)
SELECT *
FROM filled_gaps
ORDER BY name, year;
代码似乎可以运行，但我得到的是 NA 而不是“MISSING”：
name year gender country_of_birth age number_of_football_games_attended data_source
John 2010 M USA 25 3 IN ORIGINAL DATA
John 2011 M USA 26 NA BACKFILLED
John 2012 M USA 27 NA BACKFILLED
John 2013 M USA 28 NA BACKFILLED
John 2014 M USA 29 NA BACKFILLED
John 2015 M USA 30 8 IN ORIGINAL DATA
John 2016 M USA 31 NA BACKFILLED
John 2017 M USA 32 NA BACKFILLED
John 2018 M USA 33 NA BACKFILLED
John 2019 M USA 34 NA BACKFILLED
John 2020 M USA 35 12 IN ORIGINAL DATA
Maria 2012 F Brazil 22 2 IN ORIGINAL DATA
Maria 2013 F Brazil 23 NA BACKFILLED
Maria 2014 F Brazil 24 5 IN ORIGINAL DATA
Maria 2015 F Brazil 25 NA BACKFILLED
Maria 2016 F Brazil 26 NA BACKFILLED
Maria 2017 F Brazil 27 NA BACKFILLED
Maria 2018 F Brazil 28 NA BACKFILLED
Maria 2019 F Brazil 29 15 IN ORIGINAL DATA
Sofia 2018 F Spain 31 7 IN ORIGINAL DATA
Sofia 2019 F Spain 32 9 IN ORIGINAL DATA
Sofia 2020 F Spain 33 NA BACKFILLED
Sofia 2021 F Spain 34 13 IN ORIGINAL DATA
有办法解决这个问题吗？
COALESCECASECOALESCE(
CAST(m.number_of_football_games_attended AS VARCHAR),
'MISSING'
) as number_of_football_games_attended,
与：
CASE
WHEN m.number_of_football_games_attended IS NOT NULL THEN CAST(m.number_of_football_games_attended AS VARCHAR)
ELSE 'MISSING'
END as number_of_football_games_attended,