How to fill empty column with number sequence?

[Ken1]

I am going to drop a primary key from one column and create a new
column to be used as primary key in an existing database.
The old column was a date column which someone earlier though was a
good candidate for a primary key which we all know it's not.

Now I want to add a new field, i.e. called ID, with a normal number
sequence as primary key.

I have dropped the primary key, created the new column, created the
new sequence and created a trigger to generate the primary key for new
inserts.

But I don't know how to generate/update the new ID column with
sequence numbers for the already existing records.

Any ideas?

/Kenneth

[Ed prochak]

ke**@tjohoo.se (Ken1) wrote in message news:<3b******* *************** ****@posting.go ogle.com>...

I am going to drop a primary key from one column and create a new
column to be used as primary key in an existing database.
The old column was a date column which someone earlier though was a
good candidate for a primary key which we all know it's not.

Now I want to add a new field, i.e. called ID, with a normal number
sequence as primary key.

I have dropped the primary key, created the new column, created the
new sequence and created a trigger to generate the primary key for new
inserts.

But I don't know how to generate/update the new ID column with
sequence numbers for the already existing records.

Any ideas?

/Kenneth

A simple PL/SQL process can do it.
open the table for all rows,
count each row fetched,
and update the ID to the count value.

HTH,
ed

[Jia Lian Chang]

You can create a temporary trigger that fires on updates. It will be
somewhat similar to your insert trigger, e.g.
if (:new.id = -1) then
:new.id := seq_name@nextva l;

Then, just update your whole table:
update tablename
set id = -1;

Remove the temporary update trigger after that.

[boa]

Ken1 wrote:

I am going to drop a primary key from one column and create a new
column to be used as primary key in an existing database.
The old column was a date column which someone earlier though was a
good candidate for a primary key which we all know it's not.

Now I want to add a new field, i.e. called ID, with a normal number
sequence as primary key.

I have dropped the primary key, created the new column, created the
new sequence and created a trigger to generate the primary key for new
inserts.

But I don't know how to generate/update the new ID column with
sequence numbers for the already existing records.

Any ideas?
How about using rownum to generate id's for the existing rows?

alter table foo add bar number;
update foo set bar = rownum;

Remember to create the sequence with the proper "START WITH" number,
which should be max(bar) + 1.

HTH
Boa

/Kenneth

[steve]

On Fri, 13 Feb 2004 16:11:25 +0800, Ken1 wrote
(in article <3b************ **************@ posting.google. com>):

I am going to drop a primary key from one column and create a new
column to be used as primary key in an existing database.
The old column was a date column which someone earlier though was a
good candidate for a primary key which we all know it's not.

Now I want to add a new field, i.e. called ID, with a normal number
sequence as primary key.

I have dropped the primary key, created the new column, created the
new sequence and created a trigger to generate the primary key for new
inserts.

But I don't know how to generate/update the new ID column with
sequence numbers for the already existing records.

Any ideas?

/Kenneth

create a sequence.
when you add the new col to the table ,
in the default value for the col , query the sequence.

this will add a new col with the values already installed.

then for future additions add a trigger. that calls the sequence whenever the
table is inserted.

or you could be real lazy an just leave the default value in place.

[Mark C. Stock]

"steve" <[e-mail]> wrote in message
news:00******** *************** ******@news.new sguy.com...
| On Fri, 13 Feb 2004 16:11:25 +0800, Ken1 wrote
| (in article <3b************ **************@ posting.google. com>):
|
| > I am going to drop a primary key from one column and create a new
| > column to be used as primary key in an existing database.
| > The old column was a date column which someone earlier though was a
| > good candidate for a primary key which we all know it's not.
| >
| > Now I want to add a new field, i.e. called ID, with a normal number
| > sequence as primary key.
| >
| > I have dropped the primary key, created the new column, created the
| > new sequence and created a trigger to generate the primary key for new
| > inserts.
| >
| > But I don't know how to generate/update the new ID column with
| > sequence numbers for the already existing records.
| >
| > Any ideas?
| >
| > /Kenneth
|
| create a sequence.
| when you add the new col to the table ,
| in the default value for the col , query the sequence.
|
| this will add a new col with the values already installed.
|
| then for future additions add a trigger. that calls the sequence whenever
the
| table is inserted.
|
| or you could be real lazy an just leave the default value in place.
|
|
|

you can't use a sequence as a default value

from the SQL manual:

--------------------------------------------------------------------------
DEFAULT specifies a value to be assigned to the column if a subsequent
INSERT statement omits a
value for the column. The datatype of the expression must match the datatype
of the
column. The column must also be long enough to hold this expression. For the
syntax of
expr, see "Expression s" on page 5-1.
Restriction: A DEFAULT expression cannot contain references to other
columns, the
pseudocolumns CURRVAL, NEXTVAL, LEVEL, and ROWNUM, or date constants that
are not
fully specified.
--------------------------------------------------------------------------

note also that the default value can be overridden, including with an
explicit null (making default values basically useless)

but you can use the sequence in an update clause

update x
set id = x_seq.nextval;

-- mcs