I need to rename a table in Oracle but I want to be sure that any foreign keys, constraints, triggers and sequences that reference the table are updated to use the new name.
How can I be sure that I have not broken anything?
Note that I want to preserve any existing data that the table contains.
If you
ALTER TABLE old_table_name
RENAME TO new_table_name;
all the existing constraints (foreign key and other constraints) and triggers will reference the newly renamed object. Sequences have no relationship to tables so there will be no impact on the sequences (though if you mean that you are referencing the sequence in a trigger on the table, the trigger will continue to reference the same sequence after the rename). Any stored procedures that you have written that reference the old table name, however, will need to be updated to reference the new table name.
Now, while the constraints and triggers will continue to work correctly, they will retain their original names. If you have naming conventions for these objects that you want to maintain after the table name, you'd need to do more. For example, if you want a row-level before insert trigger on table FOO to be named TRG_BI_FOO and you rename the table to BAR, you'd need to alter the trigger explicitly to change its name
ALTER TRIGGER trg_bi_foo
RENAME TO trg_bi_bar;
Similarly, you'd need to rename your constraints and indexes
ALTER TABLE bar
RENAME CONSTRAINT pk_foo TO pk_bar;
It depends on what you mean by "any foreign keys, constraints, triggers and sequences that reference the table are updated to use the new name."
Any existing indexes, constraints, and triggers against the table being renamed will automatically reference the new name.
However, any naming conventions used for those objects won't automatically use the updated name. For example, if the primary key for TABLE_NAME is generally named TABLE_NAME_PK, renaming TABLE_NAME to NEW_TABLE_NAME won't automatically rename the primary key constraint to NEW_TABLE_NAME_PK.
What will need to be checked is code - packages, procedures, and functions - which referenced the old table name, as well as any triggers which referenced the old table name. Similarly, views against the old table name will break as well. The view ALL_DEPENDENCIES can help identify which of those objects need to be updated.
ALTER TABLE oldName RENAME TO newName
Will preserve the table's dependencies and data but there can always be a piece of PL/SQL that references the old name which is going to become invalid.
Related
I want to create a table with a column that references the name of a sequence I've also created. Ideally, I'd like to have a foreign key constraint that enforces this. I've tried
create table testtable (
sequence_name varchar2(128),
constraint testtableconstr
foreign key (sequence_name)
references user_sequences (sequence_name)
on delete set null
);
but I'm getting a SQL Error: ORA-01031: insufficient privileges. I suspect either this just isn't possible, or I need to add something like on update cascade. What, if anything, can I do to enforce this constraint when I insert rows into this table?
I assume you're trying to build some sort of deployment management system to keep track of your schema objects including sequences.
To do what you ask, you might explore one of the following options:
Run a report after each deployment that compares the values in your table vs. the data dictionary view, and lists any discrepancies.
Create a DDL trigger which does the insert automatically whenever a sequence is created.
Add a trigger to the table which does a query on the sequences view and raises an exception if not found.
I'm somewhat confused at what you are trying to achieve here - a sequence (effectively) only has a single value, the next number to be allocated, not all the values that have been previously allocated.
If you simply want to ensure that an attribute in the relation is populated from the sequence, then a trigger would be the right approach.
The diagram has over 40 tables, most of them have a primary key defined.
For some reason there is this one table, which has a primary key defined, but that's being ignored when I export the model to a DDL script.
This is the "offending" key (even though it's checked it is nowhere to be found on the generated DDL script):
Has anybody had the same problem? Any ideas on how to solve it?
[EDIT] This is where the key is defined:
And this is the DDL preview (yes, the primary key shows up there):
This is what happens if I try to generate the DDL for just that table (primary key still not generated):
I was finally able to identify and reproduce the problem.
It was a simple conflict of constraints.
Table MIEMBROS had a mandatory 1 to n relationship (foreign key) from another table on its primary key column and vice-versa (there was a foreign key on MIEMBROS against the other table's primary key).
This kind of relationship between two tables makes it impossible to add a record to any of them: The insert operation will return an error complaining about the foreign key restriction pointing the other table.
Anyway I realized that one of the relationships was 0 to n so I simply unchecked the "mandatory" checkbox on the foreign key definition and everything went fine.
So, in a nutshell: The Data Modeler "fails" silently if you are defining a mutual relationship (two foreign keys, one on each table against the other table) on non nullable unique columns, by not generating the primary key of one of the tables.
Such an odd behavior, if you ask me!
"This kind of relationship between two tables makes it impossible to add a record to any of them: The insert operation will return an error complaining about the foreign key restriction pointing the other table."
Actually, if you have deferred constraints, this is not impossible. The constraints can be enforced, for example, at commit time rather than immediately at insert time.
From the Data Modeler menu under File, I used Export -> DDL File. The keys appeared in the DDL, then when I went back to the diagram and did DDL Preview, it showed all the missing stuff.
Let assume that I have an empty database and I want to create two tables. There is a relationship between them. For example, one of attributes (call them b) in R is a foreign key. The definition SHOULD (database is empty, I have not executed any statement, yet) look:
create table R(
a type primary key,
b type references S(b)
);
create table S(
b type primary key
);
If I try run script with statements as above, I get error (according to line: b type references S(b)) , because S doesn't exist - that is normal and natural. It is possible to disable checking existing of tables, coresponding to foreign keys?
I know I can change order of statements. Or create tables without constraints, and add them later.
Why I'm asking. Let assume that we have many tables, with many relationships between them. Ordering tables manually will consume a lot of time, when we will want prepare kind of 'backup script'.
I know that it is possible to disable constraint as below:
alter table table_name disable constraint_name;
But that works for example with insert statement. There is a solution proper to create statement?
I'm storing multiple hierarchies within an Oracle table (standard model of parent, child, root columns) and was interested in finding out what constraints I should consider to make sure the hierarchy remains correct and valid.
I'd like to ensure that for each row the child, parent and root columns all make sense (eg. the same root can not be defined for two different hierarchies, if a child exists so does it's parent, etc.)
I know from working with linked tables that were defined with no constraints (!!) inconsistent/incomplete data always creeps in one way or another no matter how carefully the application tries to avoid it.
Now considering triggers should generally be avoided for all but the simple cases and I can't see how to code a constraint for this, is a stored procedure the only real way to go (presumably with a serializable transaction mode set)?
You can not define foreign key referring the same table of the constraint in the CREATE TABLE statement, but you could do it with an ALTER TABLE statement.
create table XXX (
key number primary key,
parent number
)
/
alter table XXX add constraint XXX_FK foreign key (parent) references XXX
/
I'm trying to create a constraint on the OE.PRODUCT_INFORMATION table which is delivered with Oracle 11g R2.
The constraint should make the PRODUCT_NAME unique.
I've tried it with the following statement:
ALTER TABLE PRODUCT_INFORMATION
ADD CONSTRAINT PRINF_NAME_UNIQUE UNIQUE (PRODUCT_NAME);
The problem is, that in the OE.PRODUCT_INFORMATION there are already product names which currently exist more than twice.
Executing the code above throws the following error:
an alter table validating constraint failed because the table has
duplicate key values.
Is there a possibility that a new created constraint won't be used on existing table data?
I've already tried the DISABLED keyword. But when I enable the constraint then I receive the same error message.
You can certainly create a constraint which will validate any newly inserted or updated records, but which will not be validated against old existing data, using the NOVALIDATE keyword, e.g.:
ALTER TABLE PRODUCT_INFORMATION
ADD CONSTRAINT PRINF_NAME_UNIQUE UNIQUE (PRODUCT_NAME)
NOVALIDATE;
If there is no index on the column, this command will create a non-unique index on the column.
If you are looking to enforce some sort of uniqueness for all future entries whilst keeping your current duplicates you cannot use a UNIQUE constraint.
You could use a trigger on the table to check the value to be inserted against the current table values and if it already exists, prevent the insert.
http://download.oracle.com/docs/cd/B19306_01/appdev.102/b14251/adfns_triggers.htm
or you could just remove the duplicate values and then enfoce your UNIQUE constraint.
EDIT: After Jonearles and Jeffrey Kemp's comments, I'll add that you can actually enable a unique constraint on a table with duplicate values present using the NOVALIDATE clause but you'd not be able to have a unique index on that constrained column.
See Tom Kyte's explanation here.
However, I would still worry about how obvious the intent was to future people who have to support the database. From a support perspective, it'd be more obvious to either remove the duplicates or use the trigger to make your intent clear.
YMMV
You can use deferrable .
ALTER TABLE PRODUCT_INFORMATION
ADD CONSTRAINT PRINF_NAME_UNIQUE UNIQUE (PRODUCT_NAME)
deferrable initially deferred NOVALIDATE;