Re: Datastructure for table

[Tim X]

Johan Andersson <johan.rejeep@gmail.com> writes:

> Hey,
>
> I have an Org-table looking like this:
> | person | car     | wife      |
> | John     | Volvo | Susan |
> | Peter    | BMW  | Greta   |
> | Stefan  | Golf    | Althea  |
>
> I want to parse this so that I get one object per row (execpt the
> header, which defines the keys). So if I for example have the first
> object, I want to be able to fetch the value of some column given the
> header key. The table is dynamic and can have any number of columns
> and the header names differ.
>
> What kind of datastructure would fit for this? I can think of a few
> that may work (Hash-table, Struct and Plist), but none that fit very
> well.
>

Well, its difficult to say what structure would work best because this
really depends on what you plan to do with the data and the relative
frequencies of the various operations you will be performing. 

Additional options you didn't mention could include 

           1 Association lists. Keys would be columns. Has the advantage
           of being able to use assoc etc to get/set values
           2 Nested lists. First sublist could be the list of column
           names. Subsequent lists could be each record. 
           4. Objects i.e. EIEIO (Enhanced Implementation of Emacs
           Interpreted Objects)

Assuming the tables will vary re: columns and column names and assuming
as these are org based tables, there not going to be that large, I would
be tempted to go with alist or nested lists. However, it really depends
on what you plan to do with the data. If your not going to be updating
the values much, nested lists could be pretty easy to implement.
However, if you need access to specific column values, alist may be
better. 

Choose the structure that makes the most frequent or most complex
operations easiest to code and maintain. 

Tim



-- 
tcross (at) rapttech dot com dot au