That is clearly understood. And how it is connected to the things against data types? You wrote that that it make sense to figure out data types compatibility. I thought the main focus in SQL variations that force different strategies of data conversion when executing statements. Explain. John
2010/3/17 Bryan Pendleton <[email protected]> > ?? ??????????? wrote: > >> No, not understood. You are focusing on JDBC compatibility from the >> database side? Or you looking for compatibility of types, not provided to >> JDBC? There are convertions to generic* types due to JDBC connectivity >> issues... SQL code would run in most cases, but data would not always be >> consistent. Consistency is a MUST, as I understood (?). Explain more >> detailed. John >> > > Hi John, > > It is true that SQL code runs in most cases, but there are many small > variations between the precise features of SQL that are supported by > the various database implementations. > > Consider, for example, the ORDER BY clause. > > Most database systems allow ORDER BY to be present in simple SELECT > statements: > > select id, firstname, lastname from employee order by lastname > > This has been part of the SQL standard for 20 years and is widely > supported. > > However, there are lots of variations on this SQL statement. For example, > is it supported to use ORDER BY on a column which is not present in the > SELECT > list, as in: > > select id from employee order by lastname > > Or, is it supported to use ORDER BY on an expression, as in: > > select lastname from employee order by lower(lastname) > > Or, is it supported to use ORDER BY in a sub-query, as in: > > select * from (select lastname from employee order by lastname) t > > Over the years, we have tracked down many such variations of SQL > statements, and > explored the issues of how Derby's support for SQL compares to the support > provided by databases such as MySQL or Postgres (or others). > > These are the sorts of compatibility issues I was particularly referring > to. > > thanks, > > bryan > >
