将非关系表拆分为多个关系表

Question

I have the following table of 60000 rows, in a MySQL 5 database, which is derived from a CSV file:

CREATE TABLE `smts_import` (
  `id` int(11) NOT NULL AUTO_INCREMENT,
  `AscCode` varchar(5) NOT NULL,
  `AscName` varchar(50) NOT NULL,
  `RgnCode` varchar(5) NOT NULL,
  `RgnName` varchar(100) NOT NULL,
  `SCode` varchar(30) NOT NULL,
  `SName` varchar(100) NOT NULL,
  `AM` int(11) NOT NULL,
  `AF` int(11) NOT NULL,
  `LG` decimal(10,4) NOT NULL,
  `LT` decimal(10,4) NOT NULL,
  PRIMARY KEY (`id`),
  UNIQUE KEY `ars_codes` (`AscCode`,`RgnCode`,`SCode`),
  KEY `s_code` (`SCode`)
);

and which contains much repeated data. SCodes are unique, and the relationship between the RgnCode and SCode fields is many-to-one, as is that between the AscCode and RgnCode fields. I want to split up (normalize) the data into three separate tables in a relational manner, such that no data are repeated:

CREATE TABLE `ascs` (
  `id` int(11) NOT NULL AUTO_INCREMENT,
  `a_code` varchar(6) NOT NULL,
  `a_name` varchar(100) NOT NULL,
  PRIMARY KEY (`id`),
  UNIQUE KEY `a_code` (`a_code`)
);

CREATE TABLE `rgns` (
  `id` int(11) NOT NULL AUTO_INCREMENT,
  `a_id` int(11) NOT NULL,
  `r_code` varchar(6) NOT NULL,
  `r_name` varchar(100) NOT NULL,
  PRIMARY KEY (`id`),
  UNIQUE KEY `a_id`+`r_code` (`a_id`, `r_code`),
  KEY `r_code` (`r_code`)
);

CREATE TABLE `sms` (
  `id` int(11) NOT NULL AUTO_INCREMENT,
  `r_id` int(11) NOT NULL,
  `s_code` varchar(16) NOT NULL,
  `s_name` varchar(100) NOT NULL,
  `a_m` int(11) NOT NULL,
  `a_f` int(11) NOT NULL,
  `lg` decimal(10,4) NOT NULL,
  `lt` decimal(10,4) NOT NULL,
  PRIMARY KEY (`id`),
  UNIQUE KEY `r_id+s_code` (`r_id`,`s_code`),
  KEY `s_code` (`s_code`)
);

where rgns . a_id is a foreign key on ascs . id , and sms . r_id is a foreign key on rgns . id .

I've created the three tables and successfully populated the first two, ascs and rgns , with unique data from the smts_import table. My problem comes when I try to populate the third table sms with just the SCode , SName , AM , AF , LG and LT fields from the smts_import table, PLUS the appropriate id from the rgns table. And here I just get lost, I've tried many variations on the following:

INSERT INTO sms (r_id, s_code, s_name, a_m, a_f, lg, lt)
SELECT DISTINCT sr.id, SCode, SName, AM, AF, LG, LT
FROM sms_import AS si, rgns AS sr, ascs AS sa
WHERE (sr.r_code = si.RgnCode)
AND (sr.a_id = sa.id)
AND (sa.a_code = si.AscCode)
ORDER BY SCode

but I just end up with too many records. How do I write this insert statement to get the appropriate fields from all records from the sms_import table, plus the correct values in the sms . r_id field from the rgns . id field?

Thanks for your help

Answer 1

"the relationship between the RgnCode and SCode fields is many-to-one" So you also get many rgns.id per SCode. If you select distinct rgns.id, SCode, etc. you will get each SCode more than once depending on how many RgnCode you have for it. I would suggest adding a id column to the sms table and create a separate table sms2rgns which contains the one-to-many relations sms.id -> rgns.id