what is the query equivalence in postgresql - mysql

Question

what is the equivalence query on postgresql?
Build database and tables
CREATE DATABASE IF NOT EXISTS lemonade;
use lemonade;
CREATE TABLE users (
id int PRIMARY KEY NOT NULL auto_increment,
name varchar(50) NOT NULL,
email varchar(50) NOT NULL
);
CREATE TABLE memories (
id int PRIMARY KEY NOT NULL auto_increment,
content varchar(50) NOT NULL,
userID int,
FOREIGN KEY (userID) REFERENCES users(id)
);
INSERT INTO users (name, email) VALUES ("Ruan", "ruan#gmail.com");
INSERT INTO users (name, email) VALUES ("Pereira", "pereira#gmail.com");
INSERT INTO memories (content, userID) VALUES ("memoria 1", 1);
INSERT INTO memories (content, userID) VALUES ("memoria 2", 1);
INSERT INTO memories (content, userID) VALUES ("memoria 3", 2);
INSERT INTO memories (content, userID) VALUES ("memoria 4", 2);
query on mysql:
select ANY_VALUE(m.id), ANY_VALUE(m.content), m.userID, ANY_VALUE(u.id), ANY_VALUE(u.name), ANY_VALUE(u.email) from memories m inner join users u on m.userID = u.id group by userID;
result:
image with result of my query
query on postgresql:
?
result: expect result equal of up image

Answer 1

The ANY_VALUE function in MySQL is used when a query is doing a GROUP BY aggregation, but columns are being selected which are not mentioned in the GROUP BY clause nor appear inside aggregate functions. In the context of your query, this means that the only columns which can be selected are the userID or another column inside an aggregate function like MAX or SUM. Technically you can also select other columns from the users table as well, assuming they are functionally dependent on the userId. As the name implies, ANY_VALUE is telling MySQL to return any value for that column from each group of records.
As far as I know/expect, the value you get from ANY_VALUE is not guaranteed to be deterministic, and so would be logically equivalent to selecting a random value for that column from each group of records. Assuming you do not care which values you get back, in Postgres you could arbitrarily just select the earliest memory for each user:
SELECT
memory_id, content, id, name, email
FROM
(
SELECT m.id AS memory_id, m.content, u.id, u.name, u.email,
ROW_NUMBER() OVER (PARTITION BY u.id ORDER BY m.id) rn
FROM memories m
INNER JOIN users u
ON m.userID = u.id
) t
WHERE rn = 1;
I think that in general you should avoid using ANY_VALUE in MySQL unless you absolutely have no other choice. A better long term solution would be to clean up the MySQL query and make it ANSI compliant. Then it would be straightforward how to port it to another database.

Answer 2

I think you need DISTINCT ON ( expression ...) clause.
New code is
select distinct on(u.id)
m.id, m.content, m.userID, u.id, u.name, u.email
from memories m
inner join users u on m.userID = u.id;
Note: DISTINCT ON is Non-Standard SQL Clause.
Form more information see

Answer 3

I just had this problem and simply solved it by using a min function. It works also on string columns. In your case it would give:
SELECT
m.userID,
min(m.id) as id,
min(m.content) as content,
min(u.id) as uid,
min(u.name) as name,
min(u.email) as email
FROM memories m
INNER JOIN users u on m.userID = u.id
GROUP BY userID;
If you were ready to use any_value, then it means that you don't care which value to return. Take the minimum one is one of those random values.