MichaelRussell
Programmer
Moving an aplication away from flat files to an RDBMS means establishing the ground rules for dealing with different offerings from Oracle, Sybase, MS. Why? Well, the aplication will be sold on into customer sites where they may have their own copy of one of these RDBMSes. This would seem to mean simply writing everything "ANSI compliant" ... but where does that get you?
In a posting elsewhere,
"The technique offered.... was first posted by well-known SQL Guru Joe Celko .... Joe participated in the creation of the ANSI SQL-92 standard, and thus is a strong supporter of ANSI SQL-92 compliant code. Joe frequently makes the point that code written to the SQL-92 standard is portable to any database that supports the standard."
However, from what I've seen so far, data-types, expressions, functions (and perhaps consequentially, results?) will differ according to which vendor-product one uses.
My question: Is there a published study /useful source showing ANSI standard versus Vendor-specific features such that one can design & use a database in a sufficiently ANSI-generic way as to escape most or all of the problems one can envisage from running one 'design-and-code' set against different vendors' RDBMSes?
It is, btw, a low-throughput OLTP system with total row-numbers coming in below 100,000 -- i.e. not big!
I have the book "The Practical SQL HAndbook" 4th ed. which does list some differences, but apparently only in areas that the book itself 'uses' in its examples. These variants seem to be mostly 'syntactical'; the absence / presence of un/desirable features, I think, isn't included.
Your helpful comments / information will be most welcome! -- Michael Russell
In a posting elsewhere,
"The technique offered.... was first posted by well-known SQL Guru Joe Celko .... Joe participated in the creation of the ANSI SQL-92 standard, and thus is a strong supporter of ANSI SQL-92 compliant code. Joe frequently makes the point that code written to the SQL-92 standard is portable to any database that supports the standard."
However, from what I've seen so far, data-types, expressions, functions (and perhaps consequentially, results?) will differ according to which vendor-product one uses.
My question: Is there a published study /useful source showing ANSI standard versus Vendor-specific features such that one can design & use a database in a sufficiently ANSI-generic way as to escape most or all of the problems one can envisage from running one 'design-and-code' set against different vendors' RDBMSes?
It is, btw, a low-throughput OLTP system with total row-numbers coming in below 100,000 -- i.e. not big!
I have the book "The Practical SQL HAndbook" 4th ed. which does list some differences, but apparently only in areas that the book itself 'uses' in its examples. These variants seem to be mostly 'syntactical'; the absence / presence of un/desirable features, I think, isn't included.
Your helpful comments / information will be most welcome! -- Michael Russell