Question

我有一个带有ID（IP地址）和因子变量（Web浏览器）的表，我需要创建另一个表，每个ID都有一个记录，以及因子变量的模式。我在考虑像SELECT ip, MODE (browser) FROM log GROUP BY ip这样的东西。

不幸的是，SQLite没有实现MODE函数，所以这不起作用。我想到构建一个包含每个浏览器计数的临时表，然后使用SELECT DISTINCT ON或RANK ()语句，但SQLite也不支持这些。

此外，在单个语句中这将是很好的，因为还有其他几个因素我的模式我也需要（并且也用相同的ID分组）。

Answer 1

要计算模式，请按browser列进行分组，获取每个组的COUNT(*)，按该值排序，然后获取具有最大值的记录。

如果您已有另一个GROUP BY，请使用相关子查询：

SELECT ip,
       (SELECT browser
        FROM log AS log2
        WHERE ip = ips.ip
        GROUP BY browser
        ORDER BY COUNT(*) DESC
        LIMIT 1)
FROM (SELECT DISTINCT ip
      FROM log) AS ips

Answer 2

有一个带有时间戳，标签和延迟的日志表。我们想查看每个标签的延迟（发送时间：ST）MODE（módusz）值，按时间戳分组。一组数据值的MODE是最常出现的值。

select L, T, avg( ST ) as MODEST, C
from (
    select L, T, ST, count( ST ) as C
    from (
            select label as L, 
                         substr( substr( timeStamp, 0, 8) || '00000000', 0, 14 ) as T, 
                         latency as ST 
            from LOG 
            order by L, T, ST 
            ) as XX 
    group by L, T, ST 
) as YY
where L || '#' || T || '#' || C in (  select L || '#' || T || '#' || max(C)
                                      from(
                                        select L, T, count( ST ) as C
                                        from (
                                          select label as L, 
                                          substr( substr( timeStamp, 0, 8) || '00000000', 0, 14 ) as T, 
                                          latency as ST
                                          from LOG 
                                        ) as XX 
                                        group by L, T, ST ) as YY
                                      group by L, T )
group by L, T, C

使用SQLite通过分组计算模式

2 个答案: