I added some rows into the table manually and also I set up the ID (auto_increment) manually. Now when I try to add new row through my app into DB table, to DB table I am getting the error , that the created ID value already exist.
How can I set manually the next ID value (for example, in table I have to IDs, so how to tell to PostgreSQL, that the next ID should be counted since the number 3)?
http://www.postgresql.org/docs/current/static/functions-sequence.html
select setval('sequence-name', <new-value>);
You can get the sequence name from the the table definition:
id | integer | not null default nextval('id_seq'::regclass)
In this case the sequence is named 'id_seq'
Edit (10x to #Glenn):
SELECT setval('id_seq', max(id)) FROM table;
I think there is a simpler way:
ALTER SEQUENCE "seq_product_id" RESTART WITH 10
Related
I have one requirement. I already have a table named WorkOrder. In this table there is a column Named WorkorderId set as primary key and identity. The next one is voucherNumber. Now I want to generate voucherNumber automatically. The condition is voucher number will not repeat. E.g., first I insert 2 rows into the table and after that I delete the 2nd entry. The next time my voucher number should be 3. Again i insert 3 more entries then after that my voucher no should be 6. Then i delete one row from this table after that my voucher number should be 7. If i delete the last row (I mean 7) then next time the voucher number should the same.
Use IDENTITY(...) when creating the column. This will make a field auto-increment its value.
You'll have to drop the column first in case that it already exists. There is no (clean) way to make this happen on already existing columns.
For further information and examples you can check out http://www.w3schools.com/sql/sql_autoincrement.asp
Edit: Sorry, I have overlooked the info that you are already using IDENTITY(...) on the PK column. Unfortunately SQL-Server can only have a single column with the IDENTITY property per table... So in this case you'll have to make use of a trigger.
This is an example:
CREATE TRIGGER CountRows
ON TestCount
AFTER UPDATE
AS
UPDATE TestCount SET Cnt = Cnt +1 WHERE ID IN (SELECT ID from inserted)
GO
In case you want to enter an IDENTIFIER to the record, it is best to use uniqueIdentifier type column. It is a string constant in the form xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx, in which each x is a hexadecimal digit in the range 0-9 or a-f. For example, 6F9619FF-8B86-D011-B42D-00C04FC964FF is a valid uniqueidentifier value.
On insertion, you can simply proceed as follows;
Insert into MyTable
(WorkorderId, WorkName) values (NewId(), 'Test')
Using this, you can be sure the Id is globally unique.
What is the best way to modify the latest added row without using a temporary table.
E.g. the table structure is
id | text | date
My current approach would be an insert with the postgresql specific command "returning id" so that I can update the table afterwards with
update myTable set date='2013-11-11' where id = lastRow
However I have the feeling that postgresql is not simply using the last row but is iterating through millions of entries until "id = lastRow" is found. How can i directly access the last added row?
update myTable date='2013-11-11' where id IN(
SELECT max(id) FROM myTable
)
Just to add to mvb13's answer (since I don't have enough points to comment directly yet) there is one word missing. Hopefully, this will save someone some time from working out the correct syntax LOL.
update myTable set date='2013-11-11' where id IN(
SELECT max(id) FROM myTable
);
Im writing a VB app that is scrubbing some data inside a DB2 database. In a few tables i want to update entire columns. For example an account number column. I am changing all account numbers to start at 1, and increment as I go down the list. Id like to be able to return both the old account number, and the new one so I can generate some kind of report I can reference so I dont lose the original values. Im updating columns as so:
DECLARE #accntnum INT
SET #accntnum = 0
UPDATE accounts
SET #accntnum = accntnum = #accntnum + 1
GO
Is there a way for me to return both the original accntnum and the new one in one table?
DB2 has a really nifty feature where you can select data from a "data change statement". This was tested on DB2 for Linux/Unix/Windows, but I think that it should also work on at least DB2 for z/OS.
For your numbering, you might considering creating a sequence, as well. Then your update would be something like:
CREATE SEQUENCE acct_seq
START WITH 1
INCREMENT BY 1
NO MAXVALUE
NO CYCLE
CACHE 24
;
SELECT accntnum AS new_acct, old_acct
FROM FINAL TABLE (
UPDATE accounts INCLUDE(old_acct INT)
SET accntnum = NEXT VALUE FOR acct_seq, old_acct = accntnum
)
ORDER BY old_acct;
The INCLUDE part creates a new column in the resulting table with the name and the data type specified, and then you can set the value in the update statement as you would any other field.
A possible solution is to add an additional column (let's call it oldaccntnum) and assign old values to that column as you do your update.
Then drop it when you no longer need it.
Here's what I'd do:
-- create a new table to track the changes.
- with columns identifying a unique key, old-vale, new-value, timestamp
-- create a trigger on the accounts table
to write the old and new values to the new table.
But, not knowing all the conditions, it may not be worth the trouble.
I have this table:
Table1:
id text
1 lala
And i want take first row and copy it, but the id 1 change to 2.
Can you help me with this problem?
A SQL table has no concept of "first" row. You can however select a row based on its characteristics. So, the following would work:
insert into Table1(id, text)
select 2, text
from Table1
where id = 1;
As another note, when creating the table, you can have the id column be auto-incremented. The syntax varies from database to database. If id were auto-incremented, then you could just do:
insert into Table1(text)
select text
from Table1
where id = 1;
And you would be confident that the new row would have a unique id.
Kate - Gordon's answer is technically correct. However, I would like to know more about why you want to do this.
If you're intent is to have the field increment with the insertion of each new row, manually setting the id column value isn't a great idea - it becomes very easy for there to be a conflict with two rows attempting to use the same id at the same time.
I would recommend using an IDENTITY field for this (MS SQL Server -- use an AUTO_INCREMENT field in MySQL). You could then do the insert as follows:
INSERT INTO Table1 (text)
SELECT text
FROM Table1
WHERE id = 1
SQL Server would automatically assign a new, unique value to the id field.
I have a table where the results are sorted using an "ORDER" column, eg:
Doc_Id Doc_Value Doc_Order
1 aaa 1
12 xxx 5
2 bbb 12
3 ccc 24
My issue is to initially set up this order column as efficiently and reusably as possible.
My initial take was to set up a scalar function that could be used as a default value when a new entry is added to the table:
ALTER FUNCTION [dbo].[Documents_Initial_Order]
( )
RETURNS int
AS
BEGIN
RETURN (SELECT ISNULL(MAX(DOC_ORDER),0) + 1 FROM dbo.Documents)
When a user wants to permute 2 documents, I can then easily switch the 2 orders.
It works nicely, but I now have a second table I need to set up the same way, and I am quite sure there is a nicer way to do it. Any idea?
Based on your comment, I think you have a very workable solution. You could make it a little more userfriendly by specifying it as a default:
alter table documents
add constraint constraint_name
default (dbo.documents_initial_order()) for doc_order
As an alternative, you could create an update trigger that copies the identity field to the doc_order field after an insert:
create trigger Doc_Trigger
on Documents
for insert
as
update d
set d.doc_order = d.doc_id
from Documents d
inner join inserted i on i.doc_id = d.doc_id
Example defining doc_id as an identity column:
create table Documents (
doc_id int identity primary key,
doc_order int,
doc_value ntext
)
It sounds like you want an identity column that you can then override once it gets it initial value. One solution would be to have two columns, once call "InitialOrder", that is an auto-increment identity column, and then a second column called doc_order that initially is set to the same value as the InitialOrder field (perhaps even as part of the insert trigger or a stored procedure if you are doing inserts that way), but give the user the ability to edit that column.
It does require an extra few bytes per record, but solves your problem, and if its of any value at all, you would have both the inital document order and the user-reset order available.
Also, I am not sure if your doc_order needs to be unique or not, but if not, you can then sort return values by doc_order and InitialOrder to ensure a consistent return sequence.
If there is no need to have any control over what that DOC_ORDER value might be, try using an identity column.