Ed Leafe wrote: > On Jan 6, 2007, at 2:07 PM, Carl Karsten wrote: > >>> But there are database systems which don't have it, so why do you >>> think >>> that this could become a generic getFields()? >> Because INFORMATION_SCHEMA is defined in the sql-92 spec. > > I logged into the public MySQL database on Paul's system,
Paul's system is not a good test. something is funky with it. for the purposes of considering INFORMATION_SCHEMA, I don't count it as a fail. > Just because it's in the database doesn't mean that generic users > will have access to it. That is not how I read the sql-92 spec. what is your take on it? (the spec, not the issue...) http://www.contrib.andrew.cmu.edu/~shadow/sql/sql1992.txt search for "21 Information Schema and Definition Schema" > In fact, I would assume that any good DBA > would hide information about the system that the user doesn't > expressly need. I think that is a bad assumption. It is assuming at least 4 things: 1, a good DBA would do it. (I have yet to see a reason why they would, and I don't think a good DBA makes changes just for the fun of it, especially something that breaks compliance.) 2. that 'need" does not include running the app. I know what you are getting at, but if the app needs to somehow get the info (which I think it does) then the DBA will give access to the info, the way the app wants to get it. 3. it can be done (I have not figured out how) 4. it could not be done to 'DESCRIBE TABLE' (which would break the current implementation.) I will also ask: how well do you expect an app to behave in a hostile environment? (hostile meaning someone deliberately did something such that the server no longer conforms to the spec.) However, it is looking like enough servers don't implement this part of the spec to rely on it. Which is surprising given it is just views on tables they do have. but a reality. I think. Carl K _______________________________________________ Post Messages to: [email protected] Subscription Maintenance: http://leafe.com/mailman/listinfo/dabo-dev
