You could create an index on the function date(), which strips the time
information.
> -Original Message-
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
> Sent: 04 August 2003 14:01
> To: PgSQL General ML
> Subject: Re: [GENERAL] Monthly table partitioning for fast purges?
>
>
> I
Good day,
Just printed out all 198 pages of the Administrators Guide which is at
http://www.postgresql.com/docs/pdf/7.3/admin-7.3.2-A4.pdf only to find the
index is full of question marks where page numbers should be!!! (Well not
all granted, but about 95% are!) Arg!
---(e
> > However you cannot currently remove a column...
>
> Oh yes you can:
>
> ALTER TABLE tbl DROP COLUMN whatever
>
> http://www.postgresql.org/docs/view.php?version=7.3&idoc=0&fil
> e=ddl-alter.html
>
> IIRC it was introduced fairly recently (7.3?)
Doh! Really should get rid of 'Practical Post
> I have a table with a bunch of records already inserted. When
> I want to add or remove fields from the, what I've done is
> produce an sql file, add the fields to the create table
> directive, and add the fields manually to each record to be
> inserted. Then I do an
>
> \i db_name
>
> whic
> I have foreign keys set up so that if, for example, a record
> in customer
> is deleted, the corresponding records in the customer_addresses table
> are also removed. However, I can't find a way of ensuring
> records in the
> address table are deleted too, given that lots of different
> tables
> // check if entry already exists
> SELECT COUNT(*) FROM tablename WHERE [cond]
> ..
> if($count >0)
> UPDATE
> else
> INSERT
>
> but this will double the hit to the database server, because
> for every
> operation I need to do SELECT COUNT(*) first. The data itself
> is not a lot,
> and
