UBE sequence and data dictionary size limit

[johndanter]

Hi List,

Does anyone know the maximum size limit E1 has on sequencing data dictionary columns?

I have a text field which is a HASH summary of an E1 Order. It's a concatenation of ITM, QTY, RLOT, LOTN, LOCN # next line and so on. (all in all about 80 bytes per line, there maybe several lines on the order)

Example for 2 lines no LOTN LOCN
4809835,333,,,#5446909,333,,,#

I then need to sequence and group by this value in order to find similar orders for batching.

Initially my field was 9000 and I could not sequence on it as the UBe would crash. So I dropped it 2000 and it now works.

So there must be a size limit?

Thanks

John

 

[alexandre.grenier]

Hi john,

Just a guess here, but could it be because JDE actually creates a BLOB field in the database when a String Data Item has a length over 2000.
it probably doesn't like doing sequencing on a blob field. is it your case?

 

[johndanter]

Hi

Yeah I think I did see a message in a log somewhere about a blob field, but I can't remember

 

[johndanter]

CNC have realised that when I build the table it's build as an NVARCHAR2 yet when they build it it turns in an NCLOB....but only on one of their FAT clients, not them all.

So I've fixed this now. But I'd still like to know the limit

I have sequenced the UBE like this ITM, QTY, RLOT, LOTN, LOCN. So the HASH field follows suit anyway

 

[Larry_Jones]

John,

it should be based on whichever DBMS has the most restrictive limit.
For example SQL Server max row size is ~ 8K (excluding BLOBS, CLOBS, etc) but both Oracle and AS/400 exceed that substantially.
Oracle and SQL Server NCHAR data type max is 4,000 characters. I think AS/400 is 32K.

Net, net JDE is "supposed" to restrict you to the minimum allowed out of the databases it supports.

 

[johndanter]

Cheers Larry

That explains why dropping it form 9000 to 2000 worked.