?

Log in

No account? Create an account
entries friends calendar profile Previous Previous Next Next
Spreadsheets and databases oh my. - Ed's journal
sobrique
sobrique
Spreadsheets and databases oh my.
Spreadsheets, were originally designed as a kind of 'page calculator'. What they did was 'operations' on arrays of numbers. This made them really good for a lot of things. Financial, or mathematical.

What they are not though, is a form of data storage.

OK, that's not entirely true. Excel, with delusions of grandeur, can actually managed data in tables, quite well. Well, better than a flat text file and notepad anyway. I mean you can sort your list, and even do lookups on values.

But it's really only much good for data of the right 'kind'. Things that you could list out.

Sadly, though, in the real world, one to one 'object -> attribute' data relations, aren't actually all that common.

But that's ok. That's why relational databases were invented.

What _they_ do, is maintain, in a sense, multiple tables. And allow links between them, such that you can have one to many, one to one, and many to many relationships.

Better yet, it lets you do this in multiple ways, so I can look up which operating system my server has installed, and THEN I can look up which servers this operating system is also installed on. All at once.

It's great. Better still, you can usually extract this kind of relational data, into a 'spreadsheet' style output - if you want a list of servers and OSes, then actually a tabulated form goes into excel quite nicely.

You can do clever things, like actually cope when you have a server cluster, that are using the same disks - many to many relationship right there.

The problem arises when one uses the spreadsheet as the primary data source. There's just some stuff that a spreadsheet just doesn't handle right.

That's even before you get into the frustration and annoyance that is relaying out and reformatting a spreadsheet, and re-inputing long hexidecimal numbers in the form of serial numbers, license keys, HBA WWNs, mac addresses, or IP addresses.

OK, so you can force it. But y'know, you could probably chop down a tree, using a masonary drill if you had the time and inclination to do so. Drill loads of holes through the trunk, and eventually it's going to fall over. But that doesn't mean doing so is particularly clever, efficient, or actually a good use of your time and materiel.
2 comments or Leave a comment
Comments
elrohana From: elrohana Date: January 11th, 2008 10:18 am (UTC) (Link)
Ha!

I am sitting here working on a spreadsheet, bashing in data that any sensible company would have in a database, so that we could just output it into pretty reports, but not this shithole, oh no, all our financial data is managed in Excel. And our other databases are all in Lotus Notes. Just shoot me now, I don't think I handle much more of this :(
darkgodfred From: darkgodfred Date: January 11th, 2008 04:57 pm (UTC) (Link)
It is of course worth noting that Lotus Notes is an e-mail client and diary schedular with delusions of adequacey. Attempting to use it as a relational database makes the decision to use Excel look sane and well balanced. And Ed has told me horror stories of management who think it's a good idea to try and implement the backbone of their ITIL system through it.

It's not even as if they don't get Access in their nice MS Office packages. It's a horrible bit of software but it will still do relational tuple data storage in a vastly superior fashion.

The worst bit is that the usual corporations that are the culprits for this behaviour are the ones who're putting down millions a year for the big, expensive database clusters with the highend DBMS software.
2 comments or Leave a comment