In version 7 of progress we had a program using word index. After upgrading to new version of progress this no longer works.
we use a utf-8 database using the default word-break table
(to be sure I put it to the zero rule: proutil database -C word-rules 0)
when we rebuild the index the result is 345 keys, while in the database are 102.544 records?
I cannot think of an obvious reason for that.
How did you upgrade? proutil conv78 etc... or dump & load? If a d&l was it a binary dump & load or a dictionary d&l? And what version did you ultimately upgrade too? 10.2 is current...
Did the db start out as UTF-8? Or was that also part of the conversion process?
We used a copy of the empty utf-8 database and loaded the df.
data was loaded with ASCII dump-files
we started with openedge 10.0A
actually we are using 10.1B02
I'm strugling with it for over a week, did rebuild index, reloaded the data in the table
I got no idea why id doesn't work.
I'm thinking of something, the data on which we put the word-index is updated through a trigger?
could that have anything to do with it?
I'm confused... your first post references version 7.
Did you disable triggers for the dump & load?
Have you opened a case with tech suppport?
first of all, we are still using progress version 7 on older hardware
we are upgrading to OpenEdge 10.1B
got some more info on my problem, and wil open a case at support (had no idea on how to explain first)
I succeeded to use the word index now. By recreating a new DB copyed form /dlc/prolang/utf/empt4.db
and loading definition file and data
I've tryed to drop the table with the word index, recreate and load data but this still doesn't work
there must be something with our DB?
I think it has something to do with the fact that we converted our DB to utf-8 iso of starting from an utf-8 empty db
(converted with proutil db-name -C convchar convert utf-8)
thanks for the reponds
I'll give some feed-back when I got result from tech support
think we got it:
by applying th work-break table
apperently db was converted to utf-8 and not based on an empty utf-8