在不更新旧行的情况下将新行插入表中

Question

Alright, i have revised the question to also include what i have so far, and what i want to do. So here goes it:

CREATE ORDER (
product_nat_id int(3) NOT NULL,
name VARCHAR(20),
PRIMARY KEY (product_nate_id)
)
INSERT INTO ORDER(product_nat_id, name) VALUES(1, 'Product 1');
INSERT INTO ORDER(product_nat_id, name) VALUES(2, 'Product 2');
INSERT INTO ORDER(product_nat_id, name) VALUES(3, 'Product 3');

CREATE TABLE INT_PRODUCT (
product_id INTEGER NOT NULL AUTO_INCREMENT,
product_nat_id INTEGER NOT NULL,
title TINYTEXT,
dateCreate TIMESTAMP CURRENT_TIMESTAMP,
CONSTRAINT INT_PRODUCT_PK PRIMARY KEY (product_id),
UNIQUE INT_PRODUCT_NK (product_nat_id));

But what i want is, whenever a record arrives with an updated value but duplicate key, i need to insert it (and not updated), but avoid duplicate constraint based on the difference in time inserted. Hope this makes sense now.

Answer 1

I would suggest the following:

1. Look up the previous record. I assume you should know what that would be

   SELECT Count(*) FROM dim WHERE recordId = '$recordid'

2. If in step 1 the records returned are larger than 0 then invalidate the 'previous' record:

   UPDATE dim SET datevalid = '$datevalue' where recordId = '$recordid' and status = 2

3. Continuing with step 1 where the ecords return in the check are larger than 0 now do the insert:

   INSERT INTO dim (recordId,field1,field2,date,status) VALUES (1,'sad','123123','2013-03-26',1)

4. If step 1 was false then just do the insert:

   INSERT INTO dim (recordId,field1,field2,date,status) VALUES (1,'sad','123123','2013-03-26',1)

I would add a status field just as an extra measure when you need to find records and distinguish between valid or invalid then you do not need to filter between dates. You can then use the status field. Also have a unique auto-increment key for every record even though the data might be the same for a set of valid and invalid records. recordId and unique key will not be the same in this case. You assign the recordId and the system will assign the unique key on the table. status = 1 is valid and status = 2 is invalid.

Hope this helps!

Answer 2

sample code of your post like as:

Insert query syntax looks like this:

INSERT INTO table (primarykeycol,col1,col2) 
VALUES (1,2,3) ON DUPLICATE KEY UPDATE col1=0, col2=col2+1

If there is already a row with primarykeycol set to 1 this query is equal to:

UPDATE table SET col1=0, col2=col2+1 WHERE primarykeycol = 1

explanation as:

1. Ordinarily to achieve the same result you would have to issue an UPDATE query, then check if there were affected rows and if not issue an INSERT query.

2. This way, you can do everything in one step – first try insert and then update if insert fails.

3. One situation for which this type of syntax is perfect is when you work with daily counters. For example, you might have a table with PostID, Date and Count columns.

4. Each day you'd have to check if you already created an entry for that day and if so increase the count column – and this can be easily substituted with one INSERT … ON DUPLICATE KEY UPDATE query.

5. Unfortunately there are some caveats. One being that when you have multiple unique indexes it will act as if you had an OR condition in WHERE clause of UPDATE query.

6. This means that multiple rows should be update, but INSERT … ON DUPLICATE KEY UPDATE will update only one row.

MySQL manual: INSERT ON DUPLICATE KEY UPDATE Syntax