Question

我有3张相关的表格。一个表有我正在寻找的行，另一个表有我需要搜索的数据，第三个表描述了我要查找的数据。我从以下查询中得到了不受欢迎的结果：

SELECT * FROM names WHERE namesKey IN ( SELECT namesKey FROM data WHERE 
    ( dataType IS 3 AND data IS 'COINCIDENCE' ) 
    AND ( dataType IS 2 AND data IS 'STATE' ) 
    AND ( dataType IS 1 AND data IS 'COUNTRY' ) );

我需要帮助根据过滤器表中的多行进行查询。我需要与第二个表中存在于多行上的键对应的行...我说的很糟糕......这是一个例子：

DROP TABLE IF EXISTS names ;
CREATE TABLE names (
    namesKey INTEGER PRIMARY KEY ASC,
    name TEXT NOT NULL
);
DROP TABLE IF EXISTS data ;
CREATE TABLE data (
    dataKey INTEGER PRIMARY KEY ASC,
    namesKey INTEGER NOT NULL,
    dataType INTEGER NOT NULL,
    data TEXT NOT NULL,
    FOREIGN KEY(namesKey) REFERENCES names(namesKey)
);
DROP TABLE IF EXISTS filter ;
CREATE TABLE filter (
    filterKey INTEGER PRIMARY KEY ASC,
    dataType INTEGER NOT NULL,
    data TEXT NOT NULL
);

INSERT INTO names( name ) VALUES ( 'name1' );
INSERT INTO names( name ) VALUES ( 'name2' );
INSERT INTO names( name ) VALUES ( 'name3' );
INSERT INTO names( name ) VALUES ( 'name4' );
INSERT INTO names( name ) VALUES ( 'name5' );
INSERT INTO names( name ) VALUES ( 'name6' );
INSERT INTO names( name ) VALUES ( 'name7' );
INSERT INTO names( name ) VALUES ( 'name8' );
INSERT INTO names( name ) VALUES ( 'name9' );

INSERT INTO data( namesKey, dataType, data ) VALUES ( 1, 1, 'COUNTRY' );
INSERT INTO data( namesKey, dataType, data ) VALUES ( 1, 2, 'STATE' );
INSERT INTO data( namesKey, dataType, data ) VALUES ( 1, 3, 'CITY' );
INSERT INTO data( namesKey, dataType, data ) VALUES ( 2, 1, 'COUNTRY' );
INSERT INTO data( namesKey, dataType, data ) VALUES ( 2, 2, 'STATE' );
INSERT INTO data( namesKey, dataType, data ) VALUES ( 2, 3, 'OTHERCITY' );
INSERT INTO data( namesKey, dataType, data ) VALUES ( 3, 1, 'COUNTRY' );
INSERT INTO data( namesKey, dataType, data ) VALUES ( 3, 2, 'STATE' );
INSERT INTO data( namesKey, dataType, data ) VALUES ( 3, 3, 'COINCIDENCE' );
INSERT INTO data( namesKey, dataType, data ) VALUES ( 4, 1, 'COUNTRY' );
INSERT INTO data( namesKey, dataType, data ) VALUES ( 4, 2, 'OTHERSTATE' );
INSERT INTO data( namesKey, dataType, data ) VALUES ( 4, 3, 'COINCIDENCE' );
INSERT INTO data( namesKey, dataType, data ) VALUES ( 5, 1, 'OTHERCOUNTRY' );
INSERT INTO data( namesKey, dataType, data ) VALUES ( 5, 2, 'RANDOM' );
INSERT INTO data( namesKey, dataType, data ) VALUES ( 5, 3, 'COINCIDENCE' );
INSERT INTO data( namesKey, dataType, data ) VALUES ( 6, 1, 'OTHERCOUNTRY' );
INSERT INTO data( namesKey, dataType, data ) VALUES ( 6, 2, 'OTHERSTATE' );
INSERT INTO data( namesKey, dataType, data ) VALUES ( 6, 3, 'COINCIDENCE' );

INSERT INTO filter( dataType, data ) VALUES ( 1, 'COUNTRY' );
INSERT INTO filter( dataType, data ) VALUES ( 2, 'STATE' );
INSERT INTO filter( dataType, data ) VALUES ( 3, 'COINCIDENCE' );

现在我需要的是能够相对可靠地运行3种不同类型的查询。

我需要搜索＆＃34;无数据＆＃34;并获得名称7,8和9
- 这个很容易：
```
SELECT * FROM names WHERE namesKey NOT IN ( SELECT namesKey FROM data ) ;
```

我需要根据数据表中的单一类型数据进行搜索

同样简单，期望的结果3,4,5和6

SELECT * FROM names WHERE 
    namesKey IN ( SELECT namesKey FROM data WHERE 
        ( dataType IS 3 AND data IS 'COINCIDENCE' ) ) 
;

我需要根据过滤表中的多行进行搜索。这个我不知道怎么做......

所需结果仅为名称3行

我可以通过

来完成

SELECT * FROM names WHERE 
    namesKey IN ( SELECT namesKey FROM data WHERE 
        ( dataType IS 3 AND data IS 'COINCIDENCE' ) )
    AND 
    namesKey IN ( SELECT namesKey FROM data WHERE 
        ( dataType IS 2 AND data IS 'STATE' ) )
    AND 
    namesKey IN ( SELECT namesKey FROM data WHERE 
        ( dataType IS 1 AND data IS 'COUNTRY' ) )
;

但这只是丑陋的首都UGH！

更糟糕的是，使用这种方法，dataType在理论上是任意大的，因此我可能最终会尝试将数十个甚至数百个子查询串起来...我甚至在尝试之前我只能编写我的字符串。把它放到SQL中。所以我正在寻找更优雅的解决方案。有什么建议？

Answer 1

如果我理解正确，您可以使用：

float:left;

的 SqlFiddleDemo

输出：

SELECT * FROM names WHERE namesKey IN (SELECT namesKey FROM data WHERE dataType IS 3 AND data IS 'COINCIDENCE' INTERSECT SELECT namesKey FROM data WHERE dataType IS 2 AND data IS 'STATE' INTERSECT SELECT namesKey FROM data WHERE dataType IS 1 AND data IS 'COUNTRY' );

或使用聚合：

╔═══════════╦═══════╗ ║ namesKey ║ name ║ ╠═══════════╬═══════╣ ║ 3 ║ name3 ║ ╚═══════════╩═══════╝

的 SqlFiddleDemo2

Answer 2

您可以直接将过滤器表与实际表连接以获取具有匹配项的行，然后仅检查所有三个搜索项匹配的那些名称键，即匹配行数与该数字相同的组所有搜索值：

SELECT namesKey
FROM data
JOIN filter USING (dataType, data)
GROUP BY namesKey
HAVING COUNT(*) = (SELECT COUNT(*) FROM filter);

然后像往常一样使用这些名称：

SELECT * 
FROM names 
WHERE namesKey IN (SELECT namesKey...);

SQLite 3 CROSS或INTERSECT复杂子查询

2 个答案: