XML Considered Harmful

[Michael F. Stemper]

On 22/09/2021 17.37, Dennis Lee Bieber wrote:
> On Wed, 22 Sep 2021 09:52:59 -0500, "Michael F. Stemper"
> <michael.stemper at gmail.com> declaimed the following:
>> On 21/09/2021 19.30, Eli the Bearded wrote:
>>> In comp.lang.python, Michael F. Stemper <michael.stemper at gmail.com> wrote:

>>>> How does CSV handle hierarchical data? For instance, I have

>>>> Can CSV files model this sort of situation?
>>>
> 	<SNIP>
>>> Yes, CSV files can model that. But it would not be my first choice of
>>> data format. (Neither would JSON.) I'd probably use XML.
>>
>> Okay. 'Go not to the elves for counsel, for they will say both no
>> and yes.' (I'm not actually surprised to find differences of opinion.)
>>
> 	You'd have to include a "level" (and/or data type if multiple objects
> can be at the same level) field (as the first field) in CSV which
> identifies how to parse the rest of the CSV data (well, technically, the
> CSV module has "parsed" it -- in terms of splitting at commas, handling
> quoted strings (which may contain commas which are not split points, etc.).
> 
> 1-generator, name
> 2-fuel, name, UOM, heat-content, price
> 2-curve, name
> 3-point, X, Y
> 3-point, X, Y
> ...
> 2-curve, name
> 3-point, X, Y
> 3-point, X, Y

This reminds me of how my (former) employer imported data models into
our systems from the 1970s until the mid-2000s. We had 80-column records
(called "card images"), that would have looked like:

FUEL0 LIGNITE    TON        13.610 043.581
UNIT1 COAL CREK1
UNIT2 ...

The specific columns for the start and end of each field on each record
were defined in a thousand-plus page document. (We modeled all of a
power system, not just economic data about generators.)

However, this doesn't seem like it would fit too well with the csv
module, since it requires a lot more logic on the part of the consuming
program.

Interesting flashback, though.

-- 
Michael F. Stemper
Deuteronomy 24:17