Kbase P14726: Why existing indexes that are modified to include new or different components are not
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
6/2/2005 |
|
Status: Unverified
GOAL:
Why existing indexes that are modified to include new or different components are
not "recreated" as the same name given to the index before its component change?
GOAL:
Why modified existing indexes have name changed in oracle dataserver?
FACT(s) (Environment):
Progress 9.1D
Oracle DataServer
FIX:
Following is what the user does that has caused the truncation of the index name:
1. Rename existing index
2. Create a new name for index
3. Delete renamed existing index
The fact that the index name is truncated is working as designed. Here is the sequence of events when reading the delta df:
.
1. An index called myindex is renamed to myoldindex but the foreign
name is still myindex.
2. A new index is added and the name is myindex. This name can not be
created in Oracle because there is already an index called myindex so
the code truncates to myinde and finds that this name is unique. We do this by taking the existing name and truncate the first character on the right. If by truncating the first character, the name is still not unique, we continue to truncate a character at a time making sure that the object name is unique. This is becasue Oracle requires the object names to be unique.
3. The delta df now has a drop index myoldindex but #2 has already been
done.
This is having to do with creating the SQL line by line from the delta df
instead of being able to read all the meta schema at once like the
ProToOra can.