Two features left - Mailing list pgsql-general
From | Jon Swinth |
---|---|
Subject | Two features left |
Date | |
Msg-id | 200211270913.42202.jswinth@atomicpc.com Whole thread Raw |
Responses |
Re: Two features left
Re: Two features left - add |
List | pgsql-general |
MS-SQLI have been using PostgreSQL on one of my projects since the beginning of the year now. Before that I used Oracle and . I am very impressed with the stability, speed, and usefulness PostgreSQL and think the 7.2.3 release will be grand. PostgreSQL wins out over the other open source DBs because it has those basic features needed for a fully formed data model such as foreign keys, transactions, and the speed to go with them. PostgreSQL is on the verge of winning big against closed source as well. What is standing in the way, in my opinion, is two features. I came to this conclusion after thinking about all the previous projects I have been involved with and how PostgreSQL could be used in place of the closed source DB in 90% of them with the following: Read locks for Foreign Key references SQL exception should not void a transaction Based on reading the email list for the past 8 months, others have voiced these issues as well. Some would say that replication and/or failover should also be on the list. However, I think interaction within the DB is more important as there is no work around in many cases. As many of you know, PostgreSQL takes a write lock on a referenced foreign key record when you update or lock a record in a transaction. This results in a great many delays and deadlocks on a high volume system that uses foreign keys. Some would say to just not use foreign keys and make the application keep things straight. Foreign keys are one of the things that attracts people to PostgreSQL, why would you want to tell them not to use them. Also, there are a lot of existing applications out there that would port themselves to use PostgreSQL but not if they have to re-write the way their software works. It is also not a safe assumption that the application will be the only thing accessing the DB. DBAs make mistakes too, and foreign keys often catch them. I have made inquires into how much it would cost to make this feature a reality to see if I could get a customer to finance it but have not received a response. The other feature is to allow transactions to continue without being forced to rollback when a SQL exception occurs. In many applications, a SQL exception is handled and an appropriate alternative generated so the transaction goes on. PostgreSQL does not support this and errors on every call made in the same transaction before calling rollback. Some people are willing and able to adjust there application code to handle this. Many people have long running transactions where this is not easily accomplished or are using a pre-existing application that they can't change. The point of this email is that I would like to be able to profess the joys and greatness of PostgreSQL to all my customers and whom ever else will listen. With these features I could do that easily.
pgsql-general by date: