数据表连接，左链接，判断多表为空时的操作GROUP_CONCAT

任务：遍历表mk_docrov_followup，当表中的某些字段为空时，判断与其相关联的：mk_docrov_lifestyle，mk_docrov_auxex，mk_docrov_medical_use，mk_docrov_referral这四张表中的所有字段，当所有字段均为空时，则unfinish++，只要这四张表中有一个字段不为空，则不能做unfinish++的操作。

一开始用死方法，if和for，最后发现在遍历表mk_docrov_followup之后至少还要做16次的if判断，太麻烦，所以探寻新大陆：

1 试着写表连接

select *

FROM 
(
`mk_docrov_followup` mkdf                          //一对一的表

)
LEFT JOIN 
  `mk_docrov_lifestyle` mklifestyle
  ON (mklifestyle.`mk_docrov_followup_id` = mkdf.`id`)           //一对多的表
  LEFT JOIN 
`mk_docrov_auxex` mkauxex
ON (mkauxex.`mk_docrov_followup_id` = mkdf.`id`)
LEFT JOIN 
  `mk_docrov_medical_use` mkmu
  ON (mkmu.`mk_docrov_followup_id` = mkdf.`id`)
  LEFT JOIN
`mk_docrov_referral` mkreferral

ON (mkreferral.`mk_docrov_followup_id` = mkdf.`id`)

成功生成一张齐全的表格

2 把需要做判断的字段选择出来

select mkdf.inspection_id AS a,mkdf.symptom_code AS b,
mkdf.symptom_other AS c,mkdf.mk_docrov_medcom AS d,mkdf.mk_docrov_adr AS e,mkdf.mk_docrov_control AS f,
mkdf.mk_docrov_hypoglycemia AS g,
mklifestyle.mk_docrov_smoke AS h,mklifestyle.mk_docrov_alcohol AS i,
mklifestyle.mk_docrov_sport AS j,mklifestyle.mk_docrov_meal AS k,
mklifestyle.mk_docrov_psyre AS l,mklifestyle.mk_docrov_combe AS m,mklifestyle.mk_docrov_salt AS n,
mklifestyle.mk_docrov_diet AS o,
mkauxex.mk_docrov_blood AS p,mkauxex.mk_docrov_other AS q,
 mkmu.mk_docrov_medical_name  AS r,mkmu.mk_docrov_method AS s,
 mkreferral.mk_docrov_referral_reason  AS t, mkreferral.mk_docrov_org  AS u


FROM 
(
`mk_docrov_followup` mkdf

)
LEFT JOIN 
  `mk_docrov_lifestyle` mklifestyle
  ON (mklifestyle.`mk_docrov_followup_id` = mkdf.`id`)
  LEFT JOIN 
`mk_docrov_auxex` mkauxex
ON (mkauxex.`mk_docrov_followup_id` = mkdf.`id`)
LEFT JOIN 
  `mk_docrov_medical_use` mkmu
  ON (mkmu.`mk_docrov_followup_id` = mkdf.`id`)
  LEFT JOIN
`mk_docrov_referral` mkreferral
ON (mkreferral.`mk_docrov_followup_id` = mkdf.`id`)

实验成功后

需要做 字符串的拼接

select GROUP_CONCAT(                                                                                        // 字符拼接CONCAT
IFNULL(mkdf.inspection_id, ''),IFNULL(mkdf.symptom_code, ''),                           //因为每个字段都可能为空，所以要ifnull，否则会出错
IFNULL(mkdf.symptom_other, ''),IFNULL(mkdf.mk_docrov_medcom, ''),
IFNULL(mkdf.mk_docrov_adr, ''),IFNULL(mkdf.mk_docrov_control, ''),
IFNULL(mkdf.mk_docrov_hypoglycemia,''),
IFNULL(mklifestyle.mk_docrov_smoke, ''),
IFNULL(mklifestyle.mk_docrov_alcohol,'') ,
IFNULL(mklifestyle.mk_docrov_sport,''),
IFNULL(mklifestyle.mk_docrov_meal,''),
IFNULL(mklifestyle.mk_docrov_psyre,''),
IFNULL(mklifestyle.mk_docrov_combe,''),
IFNULL(mklifestyle.mk_docrov_salt,''),
IFNULL(mklifestyle.mk_docrov_diet, ''),
IFNULL(mkauxex.mk_docrov_blood, ''),
IFNULL(mkauxex.mk_docrov_other, ''),
IFNULL(mkmu.mk_docrov_medical_name, ''),
IFNULL(mkmu.mk_docrov_method, '') ,
IFNULL(mkreferral.mk_docrov_referral_reason, ''), 
IFNULL(mkreferral.mk_docrov_org, '')  
)AS followupstatstic 
FROM 
(
`mk_docrov_followup` mkdf

)
LEFT JOIN 
 `mk_docrov_lifestyle` mklifestyle
 ON (mklifestyle.`mk_docrov_followup_id` = mkdf.`id`)
 LEFT JOIN 
`mk_docrov_auxex` mkauxex
ON (mkauxex.`mk_docrov_followup_id` = mkdf.`id`)
LEFT JOIN 
 `mk_docrov_medical_use` mkmu
 ON (mkmu.`mk_docrov_followup_id` = mkdf.`id`)
 LEFT JOIN
`mk_docrov_referral` mkreferral
ON (mkreferral.`mk_docrov_followup_id` = mkdf.`id`) 
GROUP BY mkdf.id                                             //主键，唯一标示的字符，id有几个，最终就会拼接出几个字符串。