tag:blogger.com,1999:blog-6411920579549337139.post352759824322472890..comments2023-12-31T05:26:17.608-08:00Comments on DATABASE DEBUNKINGS: This WeekFabian Pascalhttp://www.blogger.com/profile/01346669716885494092noreply@blogger.comBlogger1125tag:blogger.com,1999:blog-6411920579549337139.post-88803403466180533132017-08-07T14:24:30.132-07:002017-08-07T14:24:30.132-07:00About Cosette. They illustrate using an example i...About Cosette. They illustrate using an example involving UNION ALL and claim that "These two queries will produce the exact same results no matter what the input relations r and s contain, and no matter what the predicate p is.". That claim is wrong imo. If the columns in r and s do not "match" ***by ordinal position*** then there are name mismatches and in that case, the standard says that the column names in the result of the union are "implementation-dependent" and furthermore, the standard does not prescribe that an implementation must generate the same name again for the same scenario of name mismatches.Erwin Smouthttps://www.blogger.com/profile/17463579744642559811noreply@blogger.com