Mysql – Update one table from another table while sorting that table based on one column

MySQLorder-byupdate

This is the problem I'm trying to figure out in MySQL. We have an old table contains some forms submitted by our users. Somehow, the previous decision was each time a user comes to this survey, a new form will be submitted. So in the old table we easily have several rows with the same Firstname, Lastname, but different values in the other columns, and there's a timestamp column Date_Submission as well.

Now we are trying to move everything to a new table, but this time, for each person we only keep one row. And we want to keep some of the the latest old data of that user (like email, phone number, etc)

I could do the following:

update newtable, oldtable set newtable.email = oldtable.email where newtable.firstname = oldtable.firstname and newtable.lastname = oldtable.lastname;

Apparently this won't give me the "latest" old date for each person.

So I tried this one:

update newtable, oldtable set newtable.email = oldtable.email where newtable.firstname = oldtable.firstname and newtable.lastname = oldtable.lastname order by oldtable.Date_Submission;

But they MySQL will complain with:

"ERROR 1221 (HY000): Incorrect usage of UPDATE and ORDER BY".

So I'm wondering, what's the correct way to achieve this?

Best Answer

There are two options

OPTION #1 : INSERT INTO ... ON DUPLICATE KEY UPDATE

INSERT INTO newtable
SELECT * FROM oldtable
ORDER BY tmstamp_value
ON DUPLICATE KEY UPDATE email=VALUES(email), address=VALUES(address) ... ;

OPTION #2 : REPLACE INTO

REPLACE INTO newtable
SELECT * FROM oldtable
ORDER BY tmstamp_value;

It mechanically performs a DELETE and an INSERT.

CAVEAT

Notice I applied the ORDER BY against the old table. This should help get the latest data for a give LastName, FirstName.

You should also add a UNIQUE INDEX on the new table as follows:

ALTER TABLE newtable ADD UNIQUE INDEX name_ndx (lastname,firstname);

UPDATE 2013-04-08 17:13 EDT

If you are doing updates, there are two ways to do this:

1. UPDATE JOIN

UPDATE
(SELECT lastname,firstname,phone,email
FROM oldtable ORDER BY lastname,firstname) A
LEFT JOIN newtable B USING (lastname,firstname)
SET B.phone=A.phone,B.email=A.email;

2. UPDATE JOIN with GROUP BY

UPDATE
(
    SELECT BB.id,BB.phone,BB.email FROM 
    (SELECT lastname,firstname,MAX(id) id FROM oldtable GROUP BY lastname,firstname) AA
    INNER JOIN oldtable BB USING (lastname,firstname)
) A INNER JOIN newtable B USING (id)
SET B.phone=A.phone,B.email=A.email;

Give it a Try !!!

Both options will take full advantage of the UNIQUE index on name.

Related Solutions

MySQL Insert into table, where name = id, if not exist, then insert name and use that ID

Assuming you have some tables for Persons, Animals:

CREATE TABLE Person
  ( PersonID INT UNSIGNED NOT NULL AUTO_INCREMENT
  , PersonName VARCHAR(255) NOT NULL
  , CONSTRAINT Person_PK
      PRIMARY KEY (PersonID)
  , CONSTRAINT PersonName_UQ 
      UNIQUE (PersonName)
  ) ;

CREATE TABLE Animal
  ( AnimalID INT UNSIGNED NOT NULL AUTO_INCREMENT
  , AnimalName VARCHAR(255) NOT NULL
  , CONSTRAINT Animal_PK
      PRIMARY KEY (AnimalID)
  , CONSTRAINT AnimalName_UQ 
      UNIQUE (AnimalName)
  ) ;

and results:

CREATE TABLE Result
  ( RaceID INT UNSIGNED NOT NULL
  , Position INT UNSIGNED NOT NULL
  , PersonID INT UNSIGNED NOT NULL 
  , AnimalID INT UNSIGNED NOT NULL
  , Errors INT UNSIGNED NOT NULL DEFAULT 0
  , CompletionTime Time NULL DEFAULT NULL

  , CONSTRAINT Result_PK
      PRIMARY KEY (RaceID, Position)

  , CONSTRAINT Race_Person_UQ           -- assuming a Person cannot enter
      UNIQUE (RaceID, PersonID)                 -- a race twice

  , CONSTRAINT Race_Animal_UQ           -- assuming an Animal cannot enter
      UNIQUE (RaceID, AnimalID)                 -- a race twice

  , INDEX PersonID_IX (PersonID)                -- indexes for the Foreign Key
  , INDEX AnimalID_IX (AnimalID)                -- constraints:

  , CONSTRAINT Person_Result_FK     
      FOREIGN KEY (PersonID)
      REFERENCES Person (PersonID)
  , CONSTRAINT Animal_Result_FK     
      FOREIGN KEY (AnimalID)
      REFERENCES Animal (AnimalID)
  ) ;

I suggest you first bulk load the data (possibly with LOAD DATA from .txt or .csv files) in a table in MySQL (supplying race IDS. If you can't supply raceIDs but you have race names, the tables should be adjusted accordingly). You should have a Race table as well, this is just a sample procedure:

CREATE TABLE BulkData
  ( RaceID INT UNSIGNED NOT NULL
  , Position INT UNSIGNED NOT NULL
  , PersonName VARCHAR(255) NOT NULL
  , AnimalName VARCHAR(255) NOT NULL
  , Errors INT UNSIGNED NOT NULL DEFAULT 0          -- adjust datatypes according
  , CompletionTime Time NULL DEFAULT NULL           -- to your data
  ) ;

LOAD DATA INFILE '/results.txt' 
    INTO TABLE BulkData
    FIELDS TERMINATED BY ',' 
           ENCLOSED BY '"'
    LINES TERMINATED BY '\r\n' ;

Then you can manipulate them and insert them into the 2-3 tables. For Person:

INSERT INTO Person
    (PersonName)
SELECT DISTINCT
    b.PersonName
FROM
    BulkData AS b
WHERE NOT EXISTS
    ( SELECT 1
      FROM Person AS p
      WHERE p.PersonName = b.PersonName
    ) ;

Similar for Animal:

INSERT INTO Animal
    (AnimalName)
SELECT DISTINCT
    b.AnimalName
FROM
    BulkData AS b
WHERE NOT EXISTS
    ( SELECT 1
      FROM Animal AS a
      WHERE a.AnimalName = b.AnimalName
    ) ;

And then in Result:

INSERT INTO Result
    (RaceID, Position, PersonID, AnimalID, Errors, CompletionTime)
SELECT 
    b.RaceID, b.Position, p.PersonID, a.AnimalID, b.Errors, b.CompletionTime
FROM
    BulkData AS b
  JOIN
    Person AS p  ON p.PersonName = b.PersonName
  JOIN 
    Animal AS a  ON a.AnimalName = b.AnimalName
WHERE NOT EXISTS
    ( SELECT 1
      FROM Result AS r
      WHERE r.RaceID = b.RaceID
        AND r.PositionID = b.PositionID
    ) ;

If the importing results are satisfying, then you can empty the BulkData table and repeat the procedure with more files. The NOT EXISTS conditions will take care and not allow duplicates even if you try to load same data twice.

Sql-server – Update all rows from a table with random foreign key from another table

Not going to go through what you're trying to accomplish, but for your update, you're using a TOP without an ORDER BY. While SQL Server doesn't guarantee ordering for such queries, it doesn't just randomly choose, either. If you want to force a random row to be returned, you'll need to order by a row-level random number. For that, you can use ORDER BY ABS( CHECKSUM( NEWID() ) ).

Try it out against sys.objects to see the behavior of explicitly ordering randomly.

SELECT  TOP 1 name
FROM    sys.objects
ORDER BY ABS( CHECKSUM( NEWID() ) );

And good luck with whatever it is you're doing.

Edit:

@jean points out in the comments just NEWID() is all that is necessary, which is 100% correct, I just muscle-memory the ABS( CHECKSUM() ) since I tend to need row-level randomization with a % operator.

Edit 2:

So something like this, then?

UPDATE  t
    SET CategoryID = nt.CategoryID
FROM    dbo.Topic t
INNER JOIN (    SELECT  t.ID, ucl.CategoryID, 
                        Ordinal = ROW_NUMBER() OVER (
                            PARTITION BY t.ID
                            ORDER BY NEWID() )
                FROM    dbo.Topic t
                INNER JOIN dbo.UserCategoryLink ucl
                    ON  t.UserID = ucl.UserID ) nt
    ON  t.ID = nt.ID
WHERE   nt.Ordinal = 1;