[
Lists Home |
Date Index |
Thread Index
]
On Wednesday 26 February 2003 16:23, Bullard, Claude L (Len) wrote:
> Yep. I agree.
>
> Interestingly, some think it useful
> to strip the names on the way out, turn it
> into CSV, then restore these on the other
> end of the pipe. It's a closed system
> of course, but one where size makes a
> difference (RF systems).
In CSV, you don't lose the names; you just put a single copy of 'em on the
first line and don't repeat them in every line
<records>
<record>
<name>Alaric Snell</name>
<gender>M</gender>
<faveCheese>Brie</faveCheese>
</record>
<record>
<name>Sarah Pym</name>
<gender>F</gender>
<faveCheese>Feta</faveCheese>
</record>
...
</records>
->
name,gender,faveCheese
Alaric Snell,M,Brie
Sarah Pym,F,Feta
What you get with the CSV is that it's easier to type, and there are more
tools available to process it ;-) The downside is that it can only handle a
subset of XML documents: those that go root -> zero or more of the same
element -> any number of child elements -> CDATA. And if the child elements
in the penultimate level are sparse - eg, there is a large set of elements in
use in the entire file, but only a small fraction of them are used (on
average) on each line - then CSV can get a bit messy (lines with lots of
,,,,,,,,)
ABS
--
A city is like a large, complex, rabbit
- ARP
|
