Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.server -> Re: nulls in databasae design
In article <34B4DF48.AE7F1E64_at_ptd.net>, Neil Sauerwein <neils_at_ptd.net> writes:
>save a great deal of disk space is questionable. As for programming,
nulls
>are much easier to handle in the VB and Java that will be used,
but some
>back-end processes may still be written in MF COBOL, which I
assume uses null
>indicators in the same way as mainfram COBOL. Should
this drive the decision
>to use nulls or not ? Does anyone have any
opinions/experience/horror
>stories?
Thanks in advance,
Neil
Neil,
I'm working on several applications that will work in MSAccess, Essbase and
Oracle. The hardest part is keeping in mind the business rules. For instance,
MSAccess will allow NULLs in primary key fields when constructing foreign key
relationships--not true in Oracle. There were many times when I would assume a
NULL, when a developer used a blank or a '-' dash in that column. My favorite
recently is the flag columns. I'm adding constraints in Oracle to ensure that
the only values are IN ('Y','N'). Although this is a binary state column, I
may have to build indexes and such do to the convoluted nature of the logic
when the application was Access-only.
The best advice I can give is tackle on problem at a time. When you do have a workable solution, try to make it a standard for your department so that you can reduce the inventing the wheel syndrome.
--Bob. Received on Sun Jan 25 1998 - 00:00:00 CST