dbTalk Databases Forums  

Calculated value dilemma

comp.databases comp.databases


Discuss Calculated value dilemma in the comp.databases forum.



Reply
 
Thread Tools Display Modes
  #1  
Old   
Rainboy
 
Posts: n/a

Default Calculated value dilemma - 07-16-2008 , 06:00 AM






Hi all

I am designing a database for my charity (we are a local special needs
charity) to replace our existing one, and my main goals are to make
the database accurate, easy to use, and able to expand (the current
database is not!).

The database will hold contact and membership details for about 300
families, staff/volunteers and other contacts, along with event
booking (we hold about 5 a week) and payment details.

So, for each family we need to see how much they owe, plus a full list
of all their transactions. I am considering having a transaction
table, which details every payment anyone has ever made.

IDEA 1
To follow normalisation rules I should avoid having calculated values
in the table, and I should therefore simply calculate the balance
whenever I need it by querying the table. I guess as the database
grows I would have to periodically 'consolidate' the balance to
maintain efficiency.

However, I see each updated balance as a snapshot in time - when I
look back to figure out 'who paid what when', I want to be 100%
confident that the database will return the same balance it would have
returned at the time. If a rogue entry were to somehow find its way
into the table, it would mess all the balances up...

IDEA 2
As keeping historically accurate info is important, 'hard-coding' the
balance in the table seems to me like it might be the most appropriate
solution, even though it breaks the normalisation rules.

Any suggestions would be greatly appreciated!

Mark

PS I've never designed a database before, so please let me know if
I'm completely barking up the wrong tree!

Reply With Quote
  #2  
Old   
Ed Prochak
 
Posts: n/a

Default Re: Calculated value dilemma - 07-16-2008 , 12:15 PM






On Jul 16, 6:00 am, Rainboy <mark.rain... (AT) gmail (DOT) com> wrote:
Quote:
Hi all

I am designing a database for my charity (we are a local special needs
charity) to replace our existing one, and my main goals are to make
the database accurate, easy to use, and able to expand (the current
database is not!).

The database will hold contact and membership details for about 300
families, staff/volunteers and other contacts, along with event
booking (we hold about 5 a week) and payment details.

So, for each family we need to see how much they owe, plus a full list
of all their transactions. I am considering having a transaction
table, which details every payment anyone has ever made.

IDEA 1
To follow normalisation rules I should avoid having calculated values
in the table, and I should therefore simply calculate the balance
whenever I need it by querying the table. I guess as the database
grows I would have to periodically 'consolidate' the balance to
maintain efficiency.

However, I see each updated balance as a snapshot in time - when I
look back to figure out 'who paid what when', I want to be 100%
confident that the database will return the same balance it would have
returned at the time. If a rogue entry were to somehow find its way
into the table, it would mess all the balances up...
and screw up an audit, and so many other things. Have you discussed
some of the requirements with an accountant?

Quote:
IDEA 2
As keeping historically accurate info is important, 'hard-coding' the
balance in the table seems to me like it might be the most appropriate
solution, even though it breaks the normalisation rules.
Why put it in the transaction table? Why abandon normalization?

Consider creating a balance summary. Different entity. Make it a view.
If the performance of the view ever gets to be an issue, then make it
a materialized view. With the right indices there shouldn't be much of
a performance drag. I don't think you have the data volume.

Quote:
Any suggestions would be greatly appreciated!

Mark

PS I've never designed a database before, so please let me know if
I'm completely barking up the wrong tree!
You are getting overly concerned about performance before you even get
the core functionality working. With decent normalization, you should
be able to optimize later, if need be.

HTH,
ed



Reply With Quote
  #3  
Old   
Ed Prochak
 
Posts: n/a

Default Re: Calculated value dilemma - 07-16-2008 , 12:15 PM



On Jul 16, 6:00 am, Rainboy <mark.rain... (AT) gmail (DOT) com> wrote:
Quote:
Hi all

I am designing a database for my charity (we are a local special needs
charity) to replace our existing one, and my main goals are to make
the database accurate, easy to use, and able to expand (the current
database is not!).

The database will hold contact and membership details for about 300
families, staff/volunteers and other contacts, along with event
booking (we hold about 5 a week) and payment details.

So, for each family we need to see how much they owe, plus a full list
of all their transactions. I am considering having a transaction
table, which details every payment anyone has ever made.

IDEA 1
To follow normalisation rules I should avoid having calculated values
in the table, and I should therefore simply calculate the balance
whenever I need it by querying the table. I guess as the database
grows I would have to periodically 'consolidate' the balance to
maintain efficiency.

However, I see each updated balance as a snapshot in time - when I
look back to figure out 'who paid what when', I want to be 100%
confident that the database will return the same balance it would have
returned at the time. If a rogue entry were to somehow find its way
into the table, it would mess all the balances up...
and screw up an audit, and so many other things. Have you discussed
some of the requirements with an accountant?

Quote:
IDEA 2
As keeping historically accurate info is important, 'hard-coding' the
balance in the table seems to me like it might be the most appropriate
solution, even though it breaks the normalisation rules.
Why put it in the transaction table? Why abandon normalization?

Consider creating a balance summary. Different entity. Make it a view.
If the performance of the view ever gets to be an issue, then make it
a materialized view. With the right indices there shouldn't be much of
a performance drag. I don't think you have the data volume.

Quote:
Any suggestions would be greatly appreciated!

Mark

PS I've never designed a database before, so please let me know if
I'm completely barking up the wrong tree!
You are getting overly concerned about performance before you even get
the core functionality working. With decent normalization, you should
be able to optimize later, if need be.

HTH,
ed



Reply With Quote
  #4  
Old   
Ed Prochak
 
Posts: n/a

Default Re: Calculated value dilemma - 07-16-2008 , 12:15 PM



On Jul 16, 6:00 am, Rainboy <mark.rain... (AT) gmail (DOT) com> wrote:
Quote:
Hi all

I am designing a database for my charity (we are a local special needs
charity) to replace our existing one, and my main goals are to make
the database accurate, easy to use, and able to expand (the current
database is not!).

The database will hold contact and membership details for about 300
families, staff/volunteers and other contacts, along with event
booking (we hold about 5 a week) and payment details.

So, for each family we need to see how much they owe, plus a full list
of all their transactions. I am considering having a transaction
table, which details every payment anyone has ever made.

IDEA 1
To follow normalisation rules I should avoid having calculated values
in the table, and I should therefore simply calculate the balance
whenever I need it by querying the table. I guess as the database
grows I would have to periodically 'consolidate' the balance to
maintain efficiency.

However, I see each updated balance as a snapshot in time - when I
look back to figure out 'who paid what when', I want to be 100%
confident that the database will return the same balance it would have
returned at the time. If a rogue entry were to somehow find its way
into the table, it would mess all the balances up...
and screw up an audit, and so many other things. Have you discussed
some of the requirements with an accountant?

Quote:
IDEA 2
As keeping historically accurate info is important, 'hard-coding' the
balance in the table seems to me like it might be the most appropriate
solution, even though it breaks the normalisation rules.
Why put it in the transaction table? Why abandon normalization?

Consider creating a balance summary. Different entity. Make it a view.
If the performance of the view ever gets to be an issue, then make it
a materialized view. With the right indices there shouldn't be much of
a performance drag. I don't think you have the data volume.

Quote:
Any suggestions would be greatly appreciated!

Mark

PS I've never designed a database before, so please let me know if
I'm completely barking up the wrong tree!
You are getting overly concerned about performance before you even get
the core functionality working. With decent normalization, you should
be able to optimize later, if need be.

HTH,
ed



Reply With Quote
Reply




Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off



Powered by vBulletin Version 3.5.3
Copyright ©2000 - 2012, Jelsoft Enterprises Ltd.