Question

卫星表中有来自不同来源的重复卫星。

╔═════════╦══════════╦═══════╦═════════╦════════════╦══════════╦════════╗
║ _sat_id ║   name   ║ norad ║ intldes ║ un_country ║ operator ║ source ║
╠═════════╬══════════╬═══════╬═════════╬════════════╬══════════╬════════╣
║       1 ║ ISS      ║ 25544 ║ 98067A  ║ null       ║ Frank    ║ s_t    ║
║       2 ║ Int stat ║ 25544 ║ 98067A  ║ null       ║ null     ║ ucs    ║
║       3 ║ zarya    ║     0 ║ 98067A  ║ US         ║ John     ║ nasa   ║
║       4 ║ iss      ║ 25544 ║ 98067A  ║ Sovjet     ║ John     ║ celes  ║
║       5 ║ ISS      ║ 25544 ║ null    ║ Russia     ║ null     ║ other  ║
╚═════════╩══════════╩═══════╩═════════╩════════════╩══════════╩════════╝

如何按照优先级列表在MySQL中合并（而不是组concat）？例如优先级列表：

S_T
ucs
celes
NASA

所以合并的行将包含所有 1.s_t数据，所有空值都是2.ucs数据，如果仍然有空，那么3.celes数据等。

我尝试使用以下MySQL查询：

SELECT 
group_concat(`sources`) as sources,
max(`_sat_id`) as _sat_id,
max(`off_name`) as off_name,
max(`norad`) as norad,
max(`intldes`) as intldes,
max(`un_reg_country`) as un_reg_country,
max(`operator_country`) as operator_country,
max(`operator`) as operator,
max(`contractor_country`) as contractor_country,
max(`contractor`) as contractor,
max(`users`) as users,
max(`contact_info`) as contact_info,
max(`operational_status`) as operational_status,
max(`application`) as application,
max(`period`) as period,
max(`has_propulsion`) as has_propulsion,
max(`power`) as power,
max(`dry_mass`) as dry_mass,
max(`orbit_class`) as orbit_class,
max(`orbit_type`) as orbit_type,
max(`expected_life_time`) as expected_life_time,
max(`decay_date`) as decay_date,
max(`longitude`) as longitude,
max(`perigee`) as perigee,
max(`apogee`) as apogee,
max(`eccentricity`) as eccentricity,
max(`inclination`) as inclination,
max(`launch_date`) as launch_date,
max(`launch_mass`) as launch_mass,
max(`launch_site`) as launch_site,
max(`launch_vehic`) as launch_vehic,
max(`description`) as description,
group_concat(`comments`) as comments
    FROM satellite
    GROUP BY intldes

但有了这个，我不知道哪一行获得优先权。

我也尝试过使用Java，但每500行需要40秒......

提前......

Answer 1

这将返回优先级最佳匹配：

SELECT 
   intldes,
   coalesce(max(case when source = 's_t'   then `_sat_id` end)
           ,max(case when source = 'ucs'   then `_sat_id` end)
           ,max(case when source = 'celes' then `_sat_id` end)
           ,max(case when source = 'nasa'  then `_sat_id` end)) as _sat_id,
   coalesce(max(case when source = 's_t'   then `off_name ` end)
           ,max(case when source = 'ucs'   then `off_name ` end)
           ,max(case when source = 'celes' then `off_name ` end)
           ,max(case when source = 'nasa'  then `off_name ` end)) as off_name,
...
FROM satellite
GROUP BY intldes

大量的剪切和粘贴和修改，可能不是很有效。但希望这是一次性的工作。

Answer 2

我相信类似以下查询的内容可以解决问题：

SELECT
    COALESCE(s_t.off_name, ucs.off_name, celes.off_name, nasa.off_name) AS off_name
FROM (SELECT intldes FROM satellite GROUP BY intldes) all
LEFT JOIN satellite s_t ON all.intldes = s_t.intldes AND s_t.source = 's_t'
LEFT JOIN satellite ucs ON all.intldes = ucs.intldes AND ucs.source = 'ucs'
LEFT JOIN satellite celes ON all.intldes = celes.intldes AND celes.source = 'celes'
LEFT JOIN satellite nasa ON all.intldes = nasa.intldes AND nasa.source = 'nasa';

我只展示了一个属性作为示例，您可以为其余属性扩展它。

Answer 3

不完全确定结果应该是什么样子，将其放入 - 可能会有所帮助 - ROW_NUMBER函数会有帮助吗？比如ROW_NUMBER() in MySQL然后根据您的优先级选择行？

如何组合10万个MySQL行

3 个答案: