I'd like to add a NOT NULL column to a table called server. Problem is, the table already contains records. When I invoke ALTER TABLE server ADD COLUMN full_discovery_duration BIGINT NOT NULL H2 complains that full_discovery_duration may not be null. I can work around the problem by specifying DEFAULT 0 but I don't want a default valuefor future inserts. What am I supposed to do?
Should I add the column with a default and then remove DEFAULT 0 from the column definition in a subsequent statement? Is there a better way?
You can first add the column with a default value, and then set the default to null. Getting rid of the default definition is not possible however as far as I know.
An alternative is to first allow null, then set the values, and later not allow nulls.
drop table server;
create table server(id int);
insert into server values(1);
alter table server
add column
full_discovery_duration bigint;
update server set full_discovery_duration = 0;
alter table server
alter column
full_discovery_duration set not null;
While adding columns to existing tables, it should either be a nullable column or a default value must be specified. And what do you mean by removing the default? how can you remove value from a not null column?
Related
this is my PL-SQL statement
ALTER TABLE regions MODIFY (region_name VARCHAR(40) DEFAULT 'Euro') CONSTRAINT region_nn NOT NULL;
The column 'region_name' has NULL values I want to replace with 'Euro'. I get an error with this, and I'm wondering if I have the syntax wrong or if it's impossible to place a default value when adding the NOT NULL constraint and I have to do it as two separate SQL statements
Thank you for your help'
adding a constrain does not modify any existing data, it only modifies the definition of your table. Fix your data first, then add the constraint - or add the constraint with the defererred keyword and then fix the data. Either way, you'll manually have to update the data.
I want to add NOT NULL to a column, but it looks like ADD CONSTRAINT doesn’t support it. How can I add the constraint?
Cockroach does not currently allow adding these constraints to an existing table.
One workaround is to create a new table with the schema you want to use (including the NOT NULL constraint), and then migrate the data to the new table using INSERT...SELECT.
Here’s an example:
CREATE TABLE tbl2 (id INT PRIMARY KEY, col_a INT NOT NULL);
INSERT INTO tbl2 SELECT * FROM tbl1;
This assumes that tbl1 has the same number of columns with the same types and doesn’t have any NULL values in its version of col_a.
The downside of this is that it has to copy all the table data, so it is not ideal on large tables.
Another workaround would be to add a new column, with the NOT NULL constraint (which implies it would also require a DEFAULT), then use an UPDATE to set its value from the existing column, then rename the new column and drop the old one.
I am running Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 - 64bit Production, I am trying to find an equivalent to 12c's "DEFAULT ON NULL" for a table. Basically I have to create a table where the requirement is than whenever someone intentionally or inadvertently passes a NULL value, it is replaced with a DEFAULT value (in this case a NUMBER type equal to 1). Is there any easy way to do this in 11g? I know I could do a trigger on the table, but I would have to put in logic for every single column, and that seems ridiculous.
My table definition currently looks like this:
CREATE TABLE MYTABLE
(
FLAG NUMBER(1) DEFAULT 0
)
If I explicitly pass in null it WILL get stored. In that situation I was expecting the default value to be placed instead.
You can use simple 'DEFAULT' behavior for columns and don't provide anything when inserting.
CREATE TABLE table1 (
id NUMBER(8.0) NOT NULL,
col1 NUMBER(8.0) NOT NULL DEFAULT 10,
col2 NUMBER(8.0) NOT NULL DEFAULT 20,
PRIMARY KEY (id);
);
INSERT INTO table1 (id) values (123); //will result in creating a row with default values.
Also if you're using some kind of ORM you can use dynamic insert and dynamic update options. This way if you don't provide values on insert they will be set to defaults;
and on update the unchanged values will not be included into query.
This is how you can bypass explicit null values in query and defaults will apply.
I have an existing table with millions of records. Now I want to add a new column with default value. But for performances reasons I don't want to apply the new value immediately to existing records. Is there one handy way to do it?
So far, what in my mind is to work around this issue is to:
Add the new column, but do not specify a default (DDL)
Update by chunk the old rows to the default value (to avoid locking the table) (DML)
COMMIT
Alter the column to add a default (DDL)
Another option would be
Add the new column without any default value but make it accept
NULL
While inserting new record, insert NULL for this column
CREATE a AFTER INSERT trigger and update the column with it's
default value.
Something like
CREATE TRIGGER trg_update
AFTER INSERT
ON table_name
BEGIN
UPDATE table_name
SET new_column = (default_value)
WHERE Id_Column = :new.Id_Column;
END;
I am trying to add a CLOB column to a table (change a VARCHAR to a CLOB actually, but it only seems possible by adding a new column, copying and dropping the old). The column should be NOT NULL without a default value (and the table is not empty). How can I achieve this?
My original idea was to create the column with a dummy default, and change that later, but that does not seem possible:
ALTER TABLE foo RENAME COLUMN text TO text_temp;
ALTER TABLE foo ADD (
text CLOB DEFAULT '*' NOT NULL
);
UPDATE foo SET text = text_temp;
ALTER TABLE foo DROP COLUMN text_temp;
ALTER TABLE foo MODIFY (
text CLOB NOT NULL
);
-- ORA-22296: invalid ALTER TABLE option for conversion of LONG datatype to LOB
I also tried defining the column as text CLOB and adding the NOT NULL constraint later, but it gave the same error. Is there a way of doing this, short of recreating the whole table?
Does
ALTER TABLE foo MODIFY text DEFAULT NULL;
work for you?
When adding or removing a default value or a NOT NULL constraint, you don't need to specify the datatype of the column.
EDIT: To quote the Oracle documentation on ALTER TABLE:
If a column has a default value, then you can use the DEFAULT clause to change the default to NULL, but you cannot remove the default value completely. If a column has ever had a default value assigned to it, then the DATA_DEFAULT column of the USER_TAB_COLUMNS data dictionary view will always display either a default value or NULL.
This should explain why you're seeing a difference in SQL Developer.
However, I don't believe there's a significant difference between specifying DEFAULT NULL for a column and not specifying a default value. In both cases, a null value will be assumed for any column not explicitly given a value in an INSERT statement.
You are declaring the column as NOT NULL when you added it to the table, so you don't need to make it NOT NULL again. I think if you corrected the syntax in the final MODIFY clause you would still get an error, albeit a different one (precise number eludes me right now).
But what you are trying to acheive is definitely possible, with the right syntax :)
SQL> alter table t23 add ctxt clob
2 /
Table altered.
SQL> update t23 set ctxt = txt
2 /
2 rows updated.
SQL> alter table t23 modify ctxt not null
2 /
Table altered.
SQL> alter table t23 drop column txt
2 /
Table altered.
SQL> alter table t23 rename column ctxt to txt
2 /
Table altered.
SQL>