Update many rows in a table with a single statement?

Question

What is the easiest way to update many rows in a table? I have a csv file that looks like this:

|primary_key |value|
|          1 |  xyz|
|          2 |  abc|
|          3 |  def|
...

Rows with these primary keys already exist in the target table

I would like to update the target table with these values. Is there syntax so that I can write something like:

update mytable set value = ('xyz', 'abc', 'def') where primary key = (1,2,3);

Looking through the MySQL Update Reference, this site (MySQL - csv update), SO (update multiple rows, multiple db updates, update multiple rows), I suspect that the answer is "no", but I'd like to confirm that this is true.

Answer 1

First here is sample data

mysql> drop table if exists mytable;
Query OK, 0 rows affected (0.03 sec)

mysql> create table mytable
    -> (
    ->     id int not null,
    ->     value VARCHAR(255),
    ->     primary key (id)
    -> );
Query OK, 0 rows affected (0.06 sec)

mysql> insert into mytable (id) values (1),(2),(3);
Query OK, 3 rows affected (0.06 sec)
Records: 3  Duplicates: 0  Warnings: 0

mysql> select * from mytable;
+----+-------+
| id | value |
+----+-------+
|  1 | NULL  |
|  2 | NULL  |
|  3 | NULL  |
+----+-------+
3 rows in set (0.00 sec)

mysql>

Here is the new query

update mytable A inner join
(
    SELECT 1 id,'xyz' value UNION
    SELECT 2   ,'abc'       UNION
    SELECT 3   ,'def'

) B USING (id)
SET A.value = B.value;

Here is the new query executed

mysql> update mytable A inner join
    -> (
    ->     SELECT 1 id,'xyz' value UNION
    ->     SELECT 2   ,'abc'       UNION
    ->     SELECT 3   ,'def'
    -> ) B USING (id)
    -> SET A.value = B.value;
Query OK, 0 rows affected (0.06 sec)
Rows matched: 3  Changed: 0  Warnings: 0

mysql> select * from mytable;
+----+-------+
| id | value |
+----+-------+
|  1 | xyz   |
|  2 | abc   |
|  3 | def   |
+----+-------+
3 rows in set (0.00 sec)

mysql>

Answer 2

Assuming that you don't want to load the data from the CSV file into a database table and then do a correlated UPDATE,

UPDATE mytable t
   SET value = (SELECT value
                  FROM tbl_with_csv_data csv
                 WHERE csv.primary_key = t.primary_key)
 WHERE EXISTS( SELECT 1
                 FROM tbl_with_csv_data csv
                 WHERE csv.primary_key = t.primary_key)

then you should be able to use a CASE

UPDATE mytable t
   SET value = CASE WHEN primary_key = 1 THEN 'xyz'
                    WHEN primary_key = 2 THEN 'abc'
                    WHEN primary_key = 3 THEN 'def'
                    ELSE value
                END
 WHERE primary_key IN (1,2,3);