updating table rows in postgres using subquery

Question

I have this table in a postgres 8.4 database:

CREATE TABLE public.dummy
(
  address_id SERIAL,
  addr1 character(40),
  addr2 character(40),
  city character(25),
  state character(2),
  zip character(5),
  customer boolean,
  supplier boolean,
  partner boolean
  
)
WITH (
  OIDS=FALSE
);

I want to update the table. Initially i tested my query using this insert statement:

insert into address customer,supplier,partner
SELECT  
    case when cust.addr1 is not null then TRUE else FALSE end customer, 
    case when suppl.addr1 is not null then TRUE else FALSE end supplier,
    case when partn.addr1 is not null then TRUE else FALSE end partner
from (
    SELECT *
        from address) pa
    left outer join cust_original cust
        on (pa.addr1=cust.addr1 and pa.addr2=cust.addr2 and pa.city=cust.city 
            and pa.state=cust.state and substring(cust.zip,1,5) = pa.zip  )
    left outer join supp_original suppl 
        on (pa.addr1=suppl.addr1 and pa.addr2=suppl.addr2 and pa.city=suppl.city 
                and pa.state=suppl.state and pa.zip = substring(suppl.zip,1,5))
    left outer join partner_original partn
        on (pa.addr1=partn.addr1 and pa.addr2=partn.addr2 and pa.city=partn.city
                  and pa.state=partn.state and pa.zip = substring(partn.zip,1,5) )
where pa.address_id = address_id

How can I convert this to an update statement, i.e., update existing rows using values returned from a select statement?

Answer 1

There are many ways to update the rows.

When it comes to UPDATE the rows using subqueries, you can use any of these approaches.

1. Approach-1 [Using direct table reference]

UPDATE
  <table1>
SET
  customer=<table2>.customer,
  address=<table2>.address,
  partn=<table2>.partn
FROM
  <table2>
WHERE
  <table1>.address_id=<table2>.address_i;

Explanation: table1 is the table which we want to update, table2 is the table, from which we'll get the value to be replaced/updated. We are using FROM clause, to fetch the table2's data. WHERE clause will help to set the proper data mapping.

2. Approach-2 [Using SubQueries]

UPDATE
  <table1>
SET
  customer=subquery.customer,
  address=subquery.address,
  partn=subquery.partn
FROM
  (
    SELECT
      address_id, customer, address, partn
    FROM  /* big hairy SQL */ ...
  ) AS subquery
WHERE
  dummy.address_id=subquery.address_id;

Explanation: Here we are using subquerie inside the FROM clause, and giving an alias to it. So that it will act like the table.

3. Approach-3 [Using multiple Joined tables]

UPDATE
  <table1>
SET
  customer=<table2>.customer,
  address=<table2>.address,
  partn=<table2>.partn
FROM
  <table2> as t2
  JOIN <table3> as t3
  ON
    t2.id = t3.id
WHERE
  <table1>.address_id=<table2>.address_i;

Explanation: Sometimes we face the situation in that table join is so important to get proper data for the update. To do so, Postgres allows us to Join multiple tables inside the FROM clause.

4. Approach-4 [Using WITH statement]

• 4.1 [Using simple query]

WITH subquery AS (
    SELECT
      address_id,
      customer,
      address,
      partn
    FROM
      <table1>;
)
UPDATE <table-X>
SET customer = subquery.customer,
    address  = subquery.address,
    partn    = subquery.partn
FROM subquery
WHERE <table-X>.address_id = subquery.address_id;

• 4.2 [Using query with complex JOIN]

WITH subquery AS (
    SELECT address_id, customer, address, partn
    FROM
      <table1> as t1
    JOIN
      <table2> as t2
    ON
      t1.id = t2.id;
    -- You can build as COMPLEX as this query as per your need.
)
UPDATE <table-X>
SET customer = subquery.customer,
    address  = subquery.address,
    partn    = subquery.partn
FROM subquery
WHERE <table-X>.address_id = subquery.address_id;

Explanation: From Postgres 9.1, this(WITH) concept has been introduced. Using that we can make any complex queries and generate desired result. Here we are using this approach to update the table.

I hope, this would be helpful...😊

Answer 2

Another way with limit:

UPDATE tableA 
SET column = 'new value'
WHERE tableA.id in (
          SELECT tableA.id FROM tableA 
            JOIN tableB ON tableA.tableB_id = tableB.id 
           WHERE tableB.someColumn = 'XYZ' 
           LIMIT 10)

Answer 3

For PostgreSQL check https://www.postgresql.org/docs/current/sql-update.html

UPDATE tableA SET (addr1, adrr2) =
    (SELECT addr1, addr2 FROM tableB
     WHERE tableA.id = tableB.tableA_id);

Answer 4

You're after the UPDATE FROM syntax.

UPDATE 
  table T1  
SET 
  column1 = T2.column1 
FROM 
  table T2 
  INNER JOIN table T3 USING (column2) 
WHERE 
  T1.column2 = T2.column2;

References

• Code sample here: GROUP BY in UPDATE FROM clause
• And here
• Formal Syntax Specification

Answer 5

@Mayur "4.2 [Using query with complex JOIN]" with Common Table Expressions (CTEs) did the trick for me.

WITH cte AS (
SELECT e.id, e.postcode
FROM employees e
LEFT JOIN locations lc ON lc.postcode=cte.postcode
WHERE e.id=1
)
UPDATE employee_location SET lat=lc.lat, longitude=lc.longi
FROM cte
WHERE employee_location.id=cte.id;

Hope this helps... :D

Answer 6

If there are no performance gains using a join, then I prefer Common Table Expressions (CTEs) for readability:

WITH subquery AS (
    SELECT address_id, customer, address, partn
    FROM  /* big hairy SQL */ ...
)
UPDATE dummy
SET customer = subquery.customer,
    address  = subquery.address,
    partn    = subquery.partn
FROM subquery
WHERE dummy.address_id = subquery.address_id;

IMHO a bit more modern.

Answer 7

update json_source_tabcol as d
set isnullable = a.is_Nullable
from information_schema.columns as a 
where a.table_name =d.table_name 
and a.table_schema = d.table_schema 
and a.column_name = d.column_name;

Answer 8

Postgres allows:

UPDATE dummy
SET customer=subquery.customer,
    address=subquery.address,
    partn=subquery.partn
FROM (SELECT address_id, customer, address, partn
      FROM  /* big hairy SQL */ ...) AS subquery
WHERE dummy.address_id=subquery.address_id;

This syntax is not standard SQL, but it is much more convenient for this type of query than standard SQL. I believe Oracle (at least) accepts something similar.