Slightly OT: Part Numbering and some more discussion

[amedina]

Our intelligent numbering is simply to distinguish things that are off
the shelf from things we design:

8#### = design

0#### = off the shelf



It would be nice if PTC added auto numbering with special prefixes like
that. The prefix it self is irrelevant in terms of its value. It could
even be a letter or name descriptor. The actual sequencial number comes
from a number list that we request from on an oracle sheet. The oracle
keeps this number ready with all info up front. When the number or if
the number actually makes it to ERP, then the number can pull info from
this list. If it doesn't, the list at least gives some info as to what
this number used to be for. This helps in searching. I've seen several
PDM link system demos and actual systems where I see a lot of trash
numbering in actual production BOM's.



Amongst the noise or trash, I see lots of "bracket_XXX" or "guy's name_
old part number_old". Lately from one of our partners I saw some parts
prefixed with PIA_#####. I thought that was smart, it stands for "Part
in Assembly". From what I hear, this is a common term for the ERP guys,
and I had seen it before but never asked about it. So I'm wondering if
there are more standard-ly used pre-fixes and suffixes that we can use
to clean up our system.



By "fix" I'll give you an example:



82545

02113_3

82512

01235_8



Here 02113 represents a particular hose or cable or a thing with a
flexible part that can be twisted around...(flexible components suck, we
don't use that)

However, it can also be a part that can be different cut length, or it
can be a "PIA" of an assembly with the same name or related product.
Like it can be a connector in a pump assembly.



How can I tell if this part needs a WT part or not? Well if it is a
pure number like 02113 or 82512, then I know for sure it needs one. But
if I go into the parts that have _#, I have no idea without some more
investigation.



Having instead:



82545

02113_3

82512

01235_PIA

I know for sure 02113_3 needs a WT and 01235_PIA does not. Having
01235_PIA as a search result also might deter someone from using that
01235_PIA part as something they want to appear in the BOM "IE have a WT
later".



Also my automatic test script that reads a BOM and determines likely
missing WT parts, would be more accurate. Maybe an automatic MFG view
maker would be possible. This would save tons of time. Usually, creating
MFG views for releasing an assembly takes a lot of time. At least here
with windchill 8, supposedly a production ready product.



Regards,



Alfonso







Alfonso Medina O.
Mechanical Engineer

[MikeLockwood]

Prefixes and suffixes can be applied to any auto-number series (via OIR's). Can apply different prefixes / suffixes to the same series in different contexts. Can apply different prefixes / suffixes to different subtypes sharing the same series.

[rey]

We used the OIR for auto numbering some of our document soft types and it
works great except sometimes it skips a few numbers. Not sure why
yet. For example some of our numbers are
(year)(autonumber-4digits)(LL).... Sometimes It will skip from 20090020LL
to 20090025LL instead of making 20090021LL. It doesn't happen all the time
but it is annoying. Anyone has an answer?

Our WTPart soft type gets it's number from an IBA attribute (called
PartNumber) of the ProE part. At this point we are not using auto
numbering for the ProE part.

Rey

At 10:08 AM 9/15/2009, Lockwood,Mike,IRVINE,R&D wrote:
>Prefixes and suffixes can be applied to any auto-number series (via
>OIR's). Can apply different prefixes / suffixes to the same series in
>different contexts. Can apply different prefixes / suffixes to different
>subtypes sharing the same series.
>
>
>----------

[MikeLockwood]

We experienced same thing - occurs for all sequences.

Numbers are generated from each sequence by Oracle in "batches" by default in order to avoid delays when creating new records rapidly, and then discarded when Oracle is stopped / started. Can change the behavior to allocate numbers from the sequence only when needed to avoid the skipping. No impact on performance except possibly when bulk loading very large numbers of objects to Windchill.

[raimondas]

Rey,
Sequence is built Oracle functionality
once in a while it skips numbers depending o cache size.
for example if cache is set to 20 and if users will not consume those
numbers in certain period of time
then cache will expire and oracle will drop those 20 numbers
and rightaway will claim another 20

Alter the sequence cache value - and you will be fine

[wpalm_99]

Rey,
Oracle caches numbers for the sequence. If Oracle is restarted, you lose
the numbers that were cached so you skip to the next number after the lost
ones. You can disable caching with the following in sqlplus...

ALTER SEQUENCE seq_name NOCACHE;

where seq_name is the name of the sequence

You can get all the sequence names with the following...

select sequence_owner, sequence_name, cache_size from all_sequences;

Another cause for gaps in the sequence is transactions that get rolled back
after NEXTVAL is executed against the sequence.

Best Regards,
Bill Palm
Director - RAPiDS Support Center
ProductSpace Solutions, Inc.



On Tue, Sep 15, 2009 at 1:12 PM, Rey Santos <-> wrote:

> We used the OIR for auto numbering some of our document soft types and it
> works great except sometimes it skips a few numbers. Not sure why yet. For
> example some of our numbers are (year)(autonumber-4digits)(LL)....
> Sometimes It will skip from 20090020LL to 20090025LL instead of making
> 20090021LL. It doesn't happen all the time but it is annoying. Anyone has
> an answer?
>
> Our WTPart soft type gets it's number from an IBA attribute (called
> PartNumber) of the ProE part. At this point we are not using auto numbering
> for the ProE part.
>
> Rey
>
>
> At 10:08 AM 9/15/2009, Lockwood,Mike,IRVINE,R&D wrote:
>
> Prefixes and suffixes can be applied to any auto-number series (via
> OIR's). Can apply different prefixes / suffixes to the same series in
> different contexts. Can apply different prefixes / suffixes to different
> subtypes sharing the same series.
>