I have a client who is running AREV Release 3.12 on a Novell 4.x platform. I have an indexed table (STOCK.MASTER), which has a key that is a maximum of 15 characters long (the key dict item is 25 characs and left justified) (e.g. G95-02-C7E15580).

There is a Btree index on the column PART.NO, which has a maximum number of characters of 10 (the dict item is 20 characs, left justfied, data type, multi-valued and position 9) (e.g. 7E-1558).

The code is using the btee.extract as follows:

if return.stock.no then return.code=yes'

search.string='

if manuf then

  search.string=manuf' : @vm : manuf : @fm

end

if part.no then

  search.string=part.no' : @vm : part.no : @fm

end

end

return.stock.no='

return.code='

open 'dict,'stock.master' to fpd.master else fpd.master='

btree.extract( search.string, 'stock.master', fpd.master, keys, '', search.flag)

convert @vm to @fm in keys

keys.count=fieldcount(keys,@fm)

begin case….

I am not able to get the system to display the correct keys stock number(s) when searching on some part numbers (e.g. 7E-1558). The system will display the stock number for 7E-1558 if I use 7E-15; although incorrect keys (with part numbers such at 7E-1986) are also displayed. Also, I receive the same results with performing a manual SORT or SELECT command.

I am able to search on some part numbers (e.g. 9U-6917; HT-43; A200M2CX) successfully.

When I search the entire index file (!STOCK.MASTER) there are no occurrences of 7E-1558. I have removed and rebuilt the Btree indexes; copied the row(s) to another file, deleted from the original file and copied back; check all dictionary items; and everything else that I could think of.

Any help and/or suggestions will be appreciated…thanx!!!

Kimberly K. Lea

Kimberley,

(I'm an OI developer and normally skip all Arev stuff, but I accidentally clicked your message.)

I think 7E-1558 may be problematic because Arev interprets it not as a string but as a real number in scientific notation. Just in case you don't know scientific notation:

mEn   =   m     times      10 raised to the power of n

mE-n  =   m   divided by   10 raised to the power of n.

So your 7E-1558 is 0.000…(1557 zeroes here)…007, like the code name for a neutrino sized James Bond, perhaps.

- Oystein -

]]007, like the code name for a neutrino sized James Bond, perhaps

Oystein Reigem is correct. We have had this same problems with Catapillar part numbers which contain and "E". The indexing procees sees the "E" and interpets it as a scinetific notation number and handles it differently from all the rest. We haven't found a way around this particular problem. We have to educate our users of this little feature.

For a fuller understanding, may I suggest that you do a search on this discussion area to see simalar guestions and responses. One or two of mine will probably be in there.

Good Luck,

Mike

Gentlemen:

Thanx for the input!!!

k

View this thread on the forum...