Monday, September 1, 2008

Binary Superset and Normal Superset in Characterset of Oracle

We make the distinction between a binary -or TRUE- superset and a 'normal' superset. For example lets take character set A and B. Character set A is a binary superset of character set B if A describes exactly the same characters and used the same codepoint for those characters as B does. On top of that it probably also has some more characters characters at codepoints that are left free by character set B - otherwise they would be 100% the same set.
If this is the case then csscan would have flagged up a CHANGELESS conversion in the csscan output and this method will work fine. It is also possible for a character set to contain the same (and more) characters as another character set but store those characters at different code points. Although this is still a logical superset it is not a binary superset. Csscan would flag those characters up as CONVERTIBLE and we need another method of changing the character set. An example of this is (AL32)UTF8, which is a Unicode character set. It easily contains all the characters of, for example WE8ISO8859P15. However, it defines a lot of those characters at different code points. Therefore UTF8 is not classed as a binary superset of WE8ISO8859P15.

No comments:

search engine

Custom Search