Delete rows from a table with inner join conditions

There are multiple ways to do it in Oracle. Right of the bat I can think about two.

1. Straightforward, not Oracle specific (assuming shop_code is unique in table1)

   update TABLE2 t2 set SHOP_NAME = 
   (
     SELECT GET_shop_name(t1.regitry_id)  
     FROM TABLE1 T1 
     WHERE t2.shop_code = t1.shop_code
   );
   

2. Oracle specific syntax(requires inline view to be key-preserved - if shop_code is not a PK column in both tables you will very likely get "SQL Error: ORA-01779: cannot modify a column which maps to a non key-preserved table" )

   UPDATE
    (
       SELECT GET_shop_name(t1.regitry_id) AS new_name,
              t2.shop_code, t2.shop_name
       FROM TABLE2 t2 
       INNER JOIN t1 ON (t2.shop_code = t1.shop_code)
     )
   SET shop_name = new_name;
   

First, join TABLE_2 and TABLE_3 using an inner join and additionally filtering on Value_A = 'a':

SELECT
  t2.ID_1,
  t3.ID_A,
  t3.Value_A
FROM
  TABLE_2 t2
  INNER JOIN TABLE_3 t3 ON t2.ID_A = t3.ID_A
WHERE
  t3.Value_A = 'a'

This will give you the following result set:

ID_1  ID_A  Value_A
----  ----  -------
1     b     a
4     d     a

Now use the above as a derived table and join it, using an outer join this time, to TABLE_1, additionally filtering the results on Value_1 = 11:

SELECT
  t1 .ID_1,
  t23.ID_A,
  t1 .Value_1,
  t23.Value_A
FROM
  TABLE_1 t1
  LEFT JOIN
  (
    SELECT
      t2.ID_1,
      t3.ID_A,
      t3.Value_A
    FROM
      TABLE_2 t2
      INNER JOIN TABLE_3 t3 ON t2.ID_A = t3.ID_A
    WHERE
      t3.Value_A = 'a'
  ) t23 ON t1.ID_1 = t23.ID_1
WHERE
  t1.Value_1 = 11
;

That will give you the output you want:

ID_1  ID_A  Value_1  Value_A
----  ----  -------  -------
1     b     11       a
3     NULL  11       NULL

However, nesting a query is not the only way to solve your problem – you can also use a nested join, which is much more concise:

SELECT
  t1.ID_1,
  t3.ID_A,
  t1.Value_1,
  t3.Value_A
FROM
  TABLE_1 t1
  LEFT JOIN
    TABLE_2 t2
    INNER JOIN TABLE_3 t3 ON t2.ID_A = t3.ID_A AND t3.Value_A = 'a'
  ON t1.ID_1 = t2.ID_1
WHERE
  t1.Value_1 = 11
;

The last query implements exactly the same logic as the previous query: first, tables TABLE_2 and TABLE_3 are joined and the result is filtered, then it is joined to TABLE_1 and the final set is filtered again. Some people also add brackets around a nested join:

FROM
  TABLE_1 t1
  LEFT JOIN
  (
    TABLE_2 t2
    INNER JOIN TABLE_3 t3 ON t2.ID_A = t3.ID_A AND t3.Value_A = 'a'
  )
  ON t1.ID_1 = t2.ID_1

to make it clearer (perhaps both for themselves and for future maintainers) that the nested join takes place before the outer-level one, logically, although the syntax is unambiguous enough without them.

Nevertheless, many people find it confusing even with brackets, and if you find yourself struggling with it as well, there is another option – right outer join:

SELECT
  t1.ID_1,
  t3.ID_A,
  t1.Value_1,
  t3.Value_A
FROM
  TABLE_2 t2
  INNER JOIN TABLE_3 t3 ON t2.ID_A = t3.ID_A AND t3.Value_A = 'a'
  RIGHT JOIN TABLE_1 t1 ON t1.ID_1 = t2.ID_1
WHERE
  t1.Value_1 = 11
;

Again, the logical sequence of events specified by the FROM clause repeats that of the derived table solution: a join between TABLE_2 and TABLE_3, filtered, is followed by a join with TABLE_1. The different syntax does not alter the outcome and the results produced still match your requirements.