Reputation: 265
I have two different databases, say DB1 and DB2. In both DBs I have a common table called test_emp, but there is a difference in the columns of the tables:
DB1
-----
desc table1
--------------
empno number
ename varchar2
sal number
hiredate date
deptno number
DB2
-----
desc table2
--------------
empno number
ename varchar2
sal number
insert into table1 select * from table2@DBLink_Name
But here I have a problem with number of columns, target table has more columns than the source.
Here I can't specify column_names
because I am passing table as dynamic.
Can you somebody please help?
Upvotes: 2
Views: 22610
Reputation: 18808
You could look at the oracle data dictionary/metadata tables on both databases and generate the necessary sql for the insert. if i_table_name
is your input.
select list_agg(column_name,',') columns -- comma separated list of columns
from (
select column_name
from all_tab_cols
where owner = 'XYZ'
and table_name = i_table_name
intersect
select column_name
from all_tab_cols@remote_database
where owner = 'XYZ' -- could be different (?)
and table_name = i_table_name
)
You could then use this string (comma separated list of colums) to do something along the lines of..
v_sql := 'insert into ' || i_table_name || ' ' || '(' ||
v_column_list || ')' ||
' select ' || '( ' || v_column_list ||
' from ' || i_table_name || '@' || remote_db_name;
dbms_output.put_line(v_sql);
execute immediate v_sql;
I haven't tested the code. Make sure you print the insert and confirm it matches your expectations.
Upvotes: 2