我要在mysql表中插入数百万条记录。为了快速地将数据迁移到表中,我尝试使用概念。
下面是CSV文件中数据的格式(一些具有JSON数据的字段)。
"1299","1420070400000","35520","1420070400000","0","0","0","2","null","{""types"": [3]}","{""types"": [2]}","{""type"": 0, ""color"": 0, ""types"": [1], ""colors"": [2]}","null","null","null","null","{""type"": 0, ""types"": [], ""sexDrive"": 0, ""sexDrives"": [1]}","null","{""results"": [2]}","{""types"": [3]}","null","null","null","{""types"": [5]}","null","null","null","null","null","null","null","null","2020-01-30 11:48:07","2020-01-30 11:48:07"
"1300","1420070400000","48219","1420070400000","0","0","0","2","null","null","null","null","null","null","null","null","null","null","null","null","null","null","null","null","null","null","null","null","null","null","null","null","2020-01-30 11:48:07","2020-01-30 11:48:07"
"1301","1420070400000","34938","1420070400000","0","0","0","2","null","null","null","{""type"": 0, ""color"": 0, ""types"": [1], ""colors"": [3]}","null","null","{""types"": [2]}","{""pain"": 0, ""pains"": [2], ""otherPains"": [3, 4, 5, 6], ""breastSymptom"": false, ""breastSymptoms"": [1]}","null","null","null","null","null","null","null","null","null","null","null","null","null","null","null","null","2020-01-30 11:48:07","2020-01-30 11:48:07"当我试图执行代码时,它会显示错误如下-
SQLSTATE22032:<>:3140无效JSON文本:“对象成员后面缺少逗号或'}‘。列“user_day_record.notesObj”的值位于42位置
基本上,当我试图在CSV中没有JSON值的情况下执行相同的命令时,它工作得很好。但是对于文件中的JSON数据(在某些字段中),它会产生错误。
这是命令-
LOAD DATA LOCAL INFILE '$file' INTO TABLE user_day_record CHARACTER SET UTF8 FIELDS TERMINATED BY ', ' LINES TERMINATED BY '\n' 有人能帮助我在CSV文件中的字段中使用json值成功地执行命令吗?
更新-
以下是user_day_record表的结构-

发布于 2020-03-23 13:57:49
下面是加载包含JSON对象的CSV文件的解决方案。下面是我们导入时必须使用的额外选项
字段以“,”结尾,可选地以“\”括起来
escaped by '\' escaped by '\"' lines terminated by '\n'*************************** 1. row ***************************
Table: user_day_record
Create Table: CREATE TABLE `user_day_record` (
`dayRecordld` bigint(20) NOT NULL AUTO_INCREMENT,
`id` varchar(191) NOT NULL,
`userld` int(11) NOT NULL,
`daylnMillis` bigint(20) NOT NULL,
`periodStatusActive` tinyint(4) NOT NULL DEFAULT '0',
`ignoreB8T` tinyint(4) NOT NULL DEFAULT '0',
`isOCOff` tinyint(4) NOT NULL DEFAULT '0',
`version` int(11) DEFAULT NULL,
`bbt` json DEFAULT NULL,
`cervicalMucusObj` json DEFAULT NULL,
`cervicalPositionObj` json DEFAULT NULL,
`spotting` json DEFAULT NULL,
`periodStrength` json DEFAULT NULL,
`temperaturelnterferenceObj` json DEFAULT NULL,
`hygieneProductsObj` json DEFAULT NULL,
`pain` json DEFAULT NULL,
`intercourse` json DEFAULT NULL,
`ovulationTestObj` json DEFAULT NULL,
`pregnancyTestObj` json DEFAULT NULL,
`illnessObj` json DEFAULT NULL,
`medicationObj` json DEFAULT NULL,
`notesObj` json DEFAULT NULL,
`sleep` json DEFAULT NULL,
`emotionsObj` json DEFAULT NULL,
`activitiesObj` json DEFAULT NULL,
`oralContraceptives` json DEFAULT NULL,
`digestion` json DEFAULT NULL,
`pExams` json DEFAULT NULL,
`pMyBaby` json DEFAULT NULL,
`pUltrasound` json DEFAULT NULL,
`pVitamin` json DEFAULT NULL,
`pWeight` json DEFAULT NULL,
`created_at` timestamp NULL DEFAULT NULL,
`updated_at` timestamp NULL DEFAULT NULL,
PRIMARY KEY (`dayRecordld`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1
1 row in set (0.00 sec)
mysql> load data local infile '/var/root/temp_Data.csv' into table user_day_record fields terminated by ',' optionally enclosed by '\"' escaped by '\\' escaped by '\"' lines terminated by '\n';
Query OK, 3 rows affected (0.00 sec)
Records: 3 Deleted: 0 Skipped: 0 Warnings: 0
mysql>
mysql> select * from user_day_record\G
*************************** 1. row ***************************
dayRecordld: 1299
id: 1420070400000
userld: 35520
daylnMillis: 1420070400000
periodStatusActive: 0
ignoreB8T: 0
isOCOff: 0
version: 2
bbt: null
cervicalMucusObj: {"types": [3]}
cervicalPositionObj: {"types": [2]}
spotting: {"type": 0, "color": 0, "types": [1], "colors": [2]}
periodStrength: null
temperaturelnterferenceObj: null
hygieneProductsObj: null
pain: null
intercourse: {"type": 0, "types": [], "sexDrive": 0, "sexDrives": [1]}
ovulationTestObj: null
pregnancyTestObj: {"results": [2]}
illnessObj: {"types": [3]}
medicationObj: null
notesObj: null
sleep: null
emotionsObj: {"types": [5]}
activitiesObj: null
oralContraceptives: null
digestion: null
pExams: null
pMyBaby: null
pUltrasound: null
pVitamin: null
pWeight: null
created_at: 2020-01-30 11:48:07
updated_at: 2020-01-30 11:48:07
*************************** 2. row ***************************
dayRecordld: 1300
id: 1420070400000
userld: 48219
daylnMillis: 1420070400000
periodStatusActive: 0
ignoreB8T: 0
isOCOff: 0
version: 2
bbt: null
cervicalMucusObj: null
cervicalPositionObj: null
spotting: null
periodStrength: null
temperaturelnterferenceObj: null
hygieneProductsObj: null
pain: null
intercourse: null
ovulationTestObj: null
pregnancyTestObj: null
illnessObj: null
medicationObj: null
notesObj: null
sleep: null
emotionsObj: null
activitiesObj: null
oralContraceptives: null
digestion: null
pExams: null
pMyBaby: null
pUltrasound: null
pVitamin: null
pWeight: null
created_at: 2020-01-30 11:48:07
updated_at: 2020-01-30 11:48:07
*************************** 3. row ***************************
dayRecordld: 1301
id: 1420070400000
userld: 34938
daylnMillis: 1420070400000
periodStatusActive: 0
ignoreB8T: 0
isOCOff: 0
version: 2
bbt: null
cervicalMucusObj: null
cervicalPositionObj: null
spotting: {"type": 0, "color": 0, "types": [1], "colors": [3]}
periodStrength: null
temperaturelnterferenceObj: null
hygieneProductsObj: {"types": [2]}
pain: {"pain": 0, "pains": [2], "otherPains": [3, 4, 5, 6], "breastSymptom": false, "breastSymptoms": [1]}
intercourse: null
ovulationTestObj: null
pregnancyTestObj: null
illnessObj: null
medicationObj: null
notesObj: null
sleep: null
emotionsObj: null
activitiesObj: null
oralContraceptives: null
digestion: null
pExams: null
pMyBaby: null
pUltrasound: null
pVitamin: null
pWeight: null
created_at: 2020-01-30 11:48:07
updated_at: 2020-01-30 11:48:07
3 rows in set (0.00 sec)
mysql>```https://stackoverflow.com/questions/60059075
复制相似问题