[topicmapmail] xpath: stop at first match

[Carlo Moneti]

On 2006.07.31 12:23 Lars Marius Garshol wrote:
> 
> * Carlo Moneti
>> 
>> Right. But my issue is this example (any path will do):
>> /topicMap/topic[@id='$id']
> 
> If you want to do this you should use xsl:key, if you are in XSLT.
> 
> But in any case it is *very* likely that you are doing something  
> that's not really a good idea. Processing XTM with XSLT/XPath is,  
> well, not a good idea. It's extremely painful, you are pretty much  
> guaranteed not to interpret the XTM document correctly, and it really  
> brings you none of the benefits of using Topic Maps in the first  
> place, since you then have to reimplement all of the functionality  
> yourself.

I am not using XSLT (or XSL). Yes, totally non-scalable. I agree with you 
on every point.

> So I strongly recommend using a Topic Maps engine instead of working  
> directly on the XTM.

I fully agree. The application will run on a database backend (still 
exploring DB tools and datamodels). But it also reads topic maps directly 
for testing and other purposes.
> 
>> The point is that I know there is only one possible match. So I  would 
>> like the processor to stop upon finding a match and return  the value, 
>> rather than needlessly scanning the rest of the document  (in this 
>> case) or the current document fragment. I expected there  would be a 
>> syntax for doing this. I haven't found it. Is there a way?
> 
> No, there isn't. The reason is that all (or nearly all) XPath  
> implementations build a full document tree before they start  
> evaluating the XPath expression, and so by the time the XPath is  
> executed the whole document has already been read.

Right. But once the tree is built, I have to thrash through it NUMEROUS 
times. Using php's simpleXML functions, I have access to either an xpath 
construct or an iterator that treats the tree as a php multi-dimensional 
array which allows me to navigate via foreach() loop. I found that when 
I'm looking for a unique path, xpath will always scan the whole tree; 
with foreach() I can of course stop when I've found a match. The speedup 
is quite big overall. My solution is to use foreach() when searching a 
unique value, and xpath for all other searches (code is smaller, cleaner, 
and faster). It would be nice to use the xpath for both cases for cleaner 
code.

It's a shame there isn't an option for an xpath scan to stop after one 
match. Surely my case isn't that unique? Perhaps I'm missing something.

Thank you,
Carlo