[RESOLVED] What is this data called?

phpSimon

Hi all

I'm working with a MSSQL database that I didn't create. In one of the tables there's a field called 'data' that contains data a bit like that shown below. Is there a term for storing data like this, in a single row - and if so can anyone recommend some search terms so I can learn about it? I've tried Googling the apparent structure but I can't find anything much, other than some references to binary serialised data.

Many thanks

RECORD:Details  

1:  

2:1  

3:110  

4:200  

5:{ts '2012-05-31 00:00:00'}  

6:Detailed description  

7:1234-CODE 
8:5410.000000  

9:1  

10:1  

11:1  

12:1  

13:0.000000 
14:0.000000  

15:0.000000  

16:0.000000  

17:0  

18:  

19:0.000000  

20:0.000000  

ENDRECORD:9999    

RECORD:Items 
1:  

2:1  

3:100  

4:{ts '2012-05-31 00:00:00'}  

5:Description  

6:1  

7:1  

8:1  

9:CODE-34  

10:1  

11:1234.000000  

12:  

13:0  

14:  

ENDRECORD:9999

dalecosp

Seen this? Might be called "structures" or "records", also the term "High Level Analysis" or "HLA".

http://www.codebreakers-journal.com/content/view/288/27/

But don't take my word for it ... I'm a major newb. 😉

bradgrafelman

phpSimon wrote:
Is there a term for storing data like this, in a single row

Yeah, the term is "Bad Idea". 🙂

phpSimon

Thanks for your comments! I presume then that there is no official term or format for storing data like this; the authors just, for whatever reason, decided to store some data in an SQL field in a format they could parse when they wanted.

Other than the fact it appears not to be designed to any sort of standard format, why is it a bad idea?

dalecosp

I gave you 4 possible terms, and I'm thinking that the person who implemented your data storage had probably either read about these or used another system somewhere where this was commonplace. As I said, IANAE ... it almost feels like a "Big Iron" thing, though. You'll need some old warrior with a broad area of knowledge to tell you, I'd imagine.

Why is it a bad idea? Current theory seems to be that you want to offload as much of the workload to the RDMBS as possible. Grabbing a huge array from the DB and then processing it with PHP goes against that theory.

Cue the flame wars 😉

laserlight

phpSimon wrote:
I presume then that there is no official term or format for storing data like this; the authors just, for whatever reason, decided to store some data in an SQL field in a format they could parse when they wanted.

It could indeed be the use of serialisation, or it could just be a flat file database stored in a field in a relational database instead of a file.

phpSimon wrote:
Other than the fact it appears not to be designed to any sort of standard format, why is it a bad idea?

That is the main bad idea, and it is related to dalecosp's observation that "current theory seems to be that you want to offload as much of the workload to the RDMBS as possible". Basically, this is an extreme way to break the basic rule that fields in a relational database should not be multi-valued. The table is thus not even in first normal form, and essentially the relational database is reduced to being no better than a simple file.

phpSimon

Thank you both for your input, that really helps.