如何用其他表中的不同值更新多行？

Question

I have 3 tables, two of them contain date, and the last one is trying to bring data from those tables. But, two (if we don't count the id column) of its columns contain data, that only this table has.

So, what I'm trying to do, is to import data from the other tables by doing multiple updates, one column at a time.

The tables have the following structure :

test1(
  id_test1 serial NOT NULL,
  typeactemariage character varying(200) NOT NULL,
  datemariage character varying(200),
  id_lieumariage integer,
  id_conjoint integer,
  id_beaupere integer,
  id_bellemere integer,
  CONSTRAINT test1_pkey PRIMARY KEY (id_test1)
)

test2
(
  id_test2 serial NOT NULL,
  ville character varying(200),
  departement character varying(200) NOT NULL,
  CONSTRAINT test2_pkey PRIMARY KEY (id_test2)
)

test3
(
  id_test3 serial NOT NULL,
  typeacte character varying(100) NOT NULL,
  epoux character varying(100) NOT NULL,
  prenomepoux character varying(100),
  prenompereepoux character varying(100),
  nommereepoux character varying(100),
  prenommereepoux character varying(100),
  epouse character varying(100),
  prenomepouse character varying(100),
  prenompereepouse character varying(100),
  nommereepouse character varying(100),
  prenommereepouse character varying(100),
  lieu character varying(150) NOT NULL,
  dept character varying(100) NOT NULL,
  dates character varying(30),
  numvue character varying(100),
  CONSTRAINT test3_pkey PRIMARY KEY (id_test3)
)

test1 already has data in the columns typeactemariage and datemariage (and id_test1 ), and tries to bring the data from test2(id_test2) in the column (id_lieumariage) . The other columns remain empty at the moment.

This is what I wrote :

UPDATE
    test1 
SET
    id_lieumariage = l.id_test2
FROM
    test1 m,
    test2 l,
    test3 o
WHERE
    o.id_test3 = m.id_test1
    AND o.lieu = l.ville
    AND o.dept = l.departement;

The code executes, the column has data, but it only fetches with the first value seen (at row 1), which is not what I want. Is there any way to make it do a loop, to do this update on every single row?

Thanks !

---

EDIT : I'm putting pictures of my tables, and gonna try to explain better.

test1 : http://puu.sh/scVCy/8479b7e4f2.png

test2 : http://puu.sh/scVmL/901df2d74a.png

test3 : http://puu.sh/scVoz/04be151c71.png

test1 after the execution of my code : http://puu.sh/scVpC/3209b5f4a6.png

The id_lieumariage column fills itself with the first value it gets, 1 . With my code, I was expecting to get the first row with the value 1, second row with value 2, third row with value 3.

I have something like 100k+ of values that I'm gonna have to import (I only shew 3 rows here, to try to simplify my example).

So, what I'd need, is to be able to get the different id_lieumariage values for each row. Tell me if you didn't understand my explanation, I don't have a very good English.

---

EDIT 2 : I'm gonna provide the data I have :

CREATE TABLE test1()
ALTER TABLE test1 ADD COLUMN id_test1 SERIAL PRIMARY KEY, ADD COLUMN typeactemariage VARCHAR(200) NOT NULL, ADD COLUMN datemariage VARCHAR(200), ADD COLUMN id_lieumariage INTEGER, ADD COLUMN id_conjoint INTEGER, ADD COLUMN id_beaupere INTEGER, ADD COLUMN id_bellemere INTEGER;

CREATE TABLE test2()
ALTER TABLE test2 ADD COLUMN id_test2 SERIAL PRIMARY KEY, ADD COLUMN ville VARCHAR(200), ADD COLUMN departement VARCHAR(200) NOT NULL*;

CREATE TABLE test3()
ALTER TABLE test3 ADD COLUMN id_test3 SERIAL PRIMARY KEY, ADD COLUMN typeacte VARCHAR(200) NOT NULL, ADD COLUMN epoux VARCHAR(200), ADD COLUMN prenomepoux VARCHAR(200), ADD COLUMN prenompereepoux VARCHAR(200), ADD COLUMN nommereepoux VARCHAR(200), ADD COLUMN prenommereepoux VARCHAR(200), ADD COLUMN epouse VARCHAR(200), ADD COLUMN prenomepouse VARCHAR(200), ADD COLUMN nommereepouse VARCHAR(200), ADD COLUMN prenommereepouse VARCHAR(200), ADD COLUMN lieu VARCHAR(200), ADD COLUMN dept VARCHAR(200), ADD COLUMN dates VARCHAR(200), ADD COLUMN numvue VARCHAR(200);

INSERT INTO test1(typeactemariage, datemariage, id_lieumariage, id_conjoint, id_beaupere, id_bellemere)
VALUES ('Contrat de mariage', '21/11/2016', NULL, NULL, NULL, NULL),
('Contrat de mariage', '25/11/2016', NULL, NULL, NULL, NULL)
('Contrat de mariage', '11/11/2016', NULL, NULL, NULL, NULL);

INSERT INTO test2(ville, departement)
VALUES('Royan', '17'),
('Rochefort', '17'),
('Aytre', '17');

INSERT INTO test3(typeacte, epoux, prenomepoux, prenompereepoux, nommereepoux, prenommereepoux, epouse, prenomepouse, prenompereepouse, nommereepouse, prenommereepouse, lieu, dept, dates, numvue)
VALUES ('Contrat de mariage', 'DUPOND', 'Loris', 'Jacques', 'Duponne', 'Jeanne', 'BOURDOIN', 'Manon', 'Dujonni', 'Jinny', 'Royan', '17', '11/11/2016', '10/70'),
('Contrat de mariage', 'DUPONT', 'Pierre', 'Jerome', 'Dupuits', 'Jeannette', 'BOURLIAN', 'Philippine', 'Marpa', 'Elisa', 'Rochefort', '17', '21/11/2016', '10/20'),
('Contrat de mariage', 'Tintin', 'Milou', 'Haddock', 'Bond', 'Barack', 'Dutroux', 'Anna', 'Trane', 'Melissa', 'Aytre', '17', '25/11/2016', '10/25');

What I'm supposed supposed to have after the execution of my request is the following :

row(1) of test1 : 1,Contrat de mariage, 21/11/2016, 1, NULL, NULL, NULL

row(2) of test1 : 1,Contrat de mariage, 25/11/2016, 2, NULL, NULL, NULL

row(3) of test1 : 1,Contrat de mariage, 11/11/2016, 3, NULL, NULL, NULL

What I do get :

row(1) of test1 : 1,Contrat de mariage, 21/11/2016, 1, NULL, NULL, NULL

row(2) of test1 : 1,Contrat de mariage, 25/11/2016, 1, NULL, NULL, NULL

row(3) of test1 : 1,Contrat de mariage, 11/11/2016, 1, NULL, NULL, NULL

What my code is doing, is taking the first l.id_test2 it finds, and puts it on every single row. If I modify the request to have :

UPDATE
    test1 
SET
    id_lieumariage = l.id_test2
FROM
    test1 m,
    test2 l,
    test3 o
WHERE
    o.id_test3 = m.id_test1
    AND o.lieu = l.ville
    AND o.dept = l.departement
    AND id_test1 = 2

Then, what I get is :

row(1) of test1 : 1,Contrat de mariage, 21/11/2016, 1, NULL, NULL, NULL

row(2) of test1 : 1,Contrat de mariage, 25/11/2016, 1, NULL, NULL, NULL

row(3) of test1 : 1,Contrat de mariage, 11/11/2016, 1, NULL, NULL, NULL

Tell me if you need more informations !

Answer 1

My answer before edit was for MySql, because MySql tag fooled my a little bit and I tested previous query on MySql database. I don't have option to test these queries for postgresql at this moment but maybe they will help you.

UPDATE
  test1
SET
  id_lieumariage = subquery.id_test2
FROM (
  SELECT id_test2, id_test1
  FROM test1 m, test2 l, test3 o
  WHERE o.id_test3 = m.id_test1
    AND o.lieu = l.ville
    AND o.dept = l.departement) AS subquery
WHERE test1.id_test1 = subquery.id_test1;

Before you'll execute update you can test with this additional query how many rows will be affected by update:

SELECT
  count(*)
FROM (
  SELECT id_test2, id_test1
  FROM test1 m, test2 l, test3 o
  WHERE o.id_test3 = m.id_test1
    AND o.lieu = l.ville
    AND o.dept = l.departement) AS subquery
WHERE test1.id_test1 = subquery.id_test1;

select Subquery should provide joined 3 tables with all requirements and the outer update query should lookup that 'subquery' table for appropriate values to set in test1 table.

Answer 2

update test1 m
set id_lieumariage = l.id_test2
from
    test3 o
    inner join
    test2 l on o.lieu = l.ville and o.dept = l.departement
where o.id_test3 = m.id_test1
;

table test1;
 id_test1 |  typeactemariage   | datemariage | id_lieumariage | id_conjoint | id_beaupere | id_bellemere 
----------+--------------------+-------------+----------------+-------------+-------------+--------------
        1 | Contrat de mariage | 21/11/2016  |              1 |             |             |             
        2 | Contrat de mariage | 25/11/2016  |              2 |             |             |             
        3 | Contrat de mariage | 11/11/2016  |              3 |             |             |