Thread: feature request and scripting question
Hello,
Has anyone considered enabling truly globally unique identifiers (outside of the data tree)? This could be done by concatenating the MAC Address(or decimal equivalent) of Eth0 to generated OID's. Since the MAC is unique in the world (we would hope), and the OID is unique within the server installation, an ID created in this manner would be unique for posterity, or at least the forseeable future. MSSQL does something like this with GUID.
This would probably slow things down, but could be added as an optional feature to be turned off or on depending on whether the developer and/or DBA desired this functionality. I can do this myself by adding a MAC field, and changing the MAC from hex to DWORD or something for storage and access, but it would be really cool to have this feature added to the database. It would streamline this functionality a great deal for those of us with partitioned tables or server farm situations.
Now to the task at hand... Is it possible to output a script containing the create statements for all structures in the current database, including permissions, procedures, etc.? I need to put the same database on multiple servers in a farm environment so need to be able to duplicate the structures. While it is true that I keep all of my scripts that I have ever added, backed up in a file, I am on a team with other developers and cannot be sure that I have every change that has ever been made to the database...
An observation:
I noticed that MySQL people changed the "simultaneous connections" to "simultaneous connections(installation default)" in the crash me test results... It was a bit misleading before... still is if you think about it. They ought to do something like default(max) to display like this in simultaneous connections it would make for a true "comparison":
PostGreSQL 32(1024)
MySQL 101(101)
If it is the installation default, it does not belong under the heading "Other Limits". This implies that 32 connections is one of PostGreSQL's "limitations".
These things are supposed to be provided: (in their own words) "... by TCX so one can get the real limitations from the database server (not the information from sales managers!). "... well if you have to go further to find out that mysql is advertising it's max against PostGreSQL's default, the results are not really informative, and decidedly sales hype. Don't list the default, under the heading "Other Limits". Call it "Default Configured Limits" or something qualified. 32 connections is simply not one of the "Other" limits. It is no limit at all. Better yet, put the real limits in these results!!!
Ok. I have gotten out my ire for the overzealous MySQL pundits. I feel better. I posted something to MySQL's general list about this as well... misleading information and misrepresentation is the best thing that you can do to harm the image of your product(and the people that develop with it). Maybe they will learn that lesson the hard way.
thx,
Neil P Davis
> Has anyone considered enabling truly globally unique identifiers (outside of the data tree)? This could be done by concatenating the MAC Address(or decimal equivalent) of Eth0 to generated OID's. MAC Ethernet addresses aren't guranteed to be unique. On the other hand, what if a host has no Ethernet NICs? More than one? > I can do this myself by adding a MAC field, and changing the MAC from hex to DWORD or something for storage and access DWORD for Ethernet MAC address? Beg your pardon? > It would streamline this functionality a great deal for those of us with partitioned tables or server farm situations. Visit one of InterBase/IBPhoenix/whatever-their-name-now-is sites and try finding papers on primary keys in distributed environments. Usefull to study and follow. -- Sniper's rifle is an extension of his eye. He kills with his injurious vision. JM
I have a database of questions and I would like to be able to pull them out randomly from the database using query. I'm thinking of something along the lines of: SELECT random() as RND,question FROM Problems ORDER BY RND; Where random() is a user defined or built in function that returns a random number. Is this possible? Is there a better way to do this on the database side. I could do it through code in the application, I'd just prefer not to. Shawn