RE: [SQL] filling an empty table w/ defaults defined - Mailing list pgsql-sql

From Jackson, DeJuan
Subject RE: [SQL] filling an empty table w/ defaults defined
Date
Msg-id D05EF808F2DFD211AE4A00105AA1B5D20DCC2B@cpsmail
Whole thread Raw
List pgsql-sql
The problem is that the datetime entry is generating the same time values in
the transaction.  You do not want that as your primary key.  You can still
create an index on the value, but if two (or more) rows are ever inserted in
the same transaction or even at the exact same time your primary key will be
invalid.
If you can't recreate the table with another primary key the just drop the
index.  You can create a new index with the CREATE INDEX statement.

> Is there a way to "turn off" a field (insertdt datetime primary key
> default 
> text 'now') as I populate a table?  I have a populated table, and a mirror
> 
> image of its structure except with the field shown above added.  I'd like 
> to insert data into the new table from the old, but postgres is telling me
> 
> it cannot insert a duplicate key into a unique index - I figure it is 
> trying to use the same date/time value for 'now'.  is there any other way?
> 
> jt
> 


pgsql-sql by date:

Previous
From: JT Kirkpatrick
Date:
Subject: incrementing by the second
Next
From: "Jackson, DeJuan"
Date:
Subject: RE: [SQL] incrementing by the second