xsl-list
[Top] [All Lists]

Re: [xsl] Running the same transformation on many input files, optimisation possible?

2019-12-15 16:50:14
Indeed, I overlooked the possibility of having the controlling loop written in 
XSLT 3.0 and executing the per-transformation code using fn:transform(), which 
gets over the problems of changing existing XSLT code if it relies on global 
variables.

Michael Kay
Saxonica

On 15 Dec 2019, at 22:35, Dimitre Novatchev dnovatchev(_at_)gmail(_dot_)com 
<xsl-list-service(_at_)lists(_dot_)mulberrytech(_dot_)com> wrote:

Note that there's a double overhead here: firstly you're bringing up a new 
Java VM for each transformation, 
and secondly you're recompiling the stylesheet for each transformation.  

Isn't the meaning of the    
    "cache" : true()

key-value pair in the $options argument of `fn:transform()` exactly to 
compile the stylesheet only once and to reuse the compiled result on any time 
in the future the same stylesheet node or stylesheet text is passed?

From the Spec (https://www.w3.org/TR/xpath-functions-31/#func-transform 
<https://www.w3.org/TR/xpath-functions-31/#func-transform>):

cache 1.0, 2.0, 3.0   This option has no effect on the result of the 
transformation but may affect efficiency. The value true indicates an 
expectation that the same stylesheet is likely to be used for more than one 
transformation; the value false indicates an expectation that the stylesheet 
will be used once only.
Type: xs:boolean
Default: true()


Thanks,
Dimitre


On Sun, Dec 15, 2019 at 2:12 PM Michael Kay mike(_at_)saxonica(_dot_)com 
<mailto:mike(_at_)saxonica(_dot_)com> 
<xsl-list-service(_at_)lists(_dot_)mulberrytech(_dot_)com 
<mailto:xsl-list-service(_at_)lists(_dot_)mulberrytech(_dot_)com>> wrote:
Note that there's a double overhead here: firstly you're bringing up a new 
Java VM for each transformation, and secondly you're recompiling the 
stylesheet for each transformation.

You can avoid the Java loading overhead by using ant or XProc, but I'm not 
sure either of them will avoid the overhead of recompiling the stylesheet; 
though if you use a a recent Saxon version, you could achieve that by 
reloading the stylesheet from a pre-compiled SEF (stylesheet export file).

You could write your own Java application to control the process, invoking 
Saxon via the JAXP or s9api APIs - both allow you to compile a stylesheet 
once and execute it repeatedly.

You might be able to write the control loop in XSLT, for example by using the 
collection() function, or functions in the EXPath file module. However, this 
could require stylesheet changes if your XSLT code binds global variables to 
values derived from the source document.

In very simple cases you can take advantage of the fact that the -s option 
for the Saxon command line can be a directory, in which case all the input 
files are transformed to corresponding files in the -o directory.

Michael Kay
Saxonica

On 15 Dec 2019, at 09:03, Trevor Nicholls 
trevor(_at_)castingthevoid(_dot_)com 
<mailto:trevor(_at_)castingthevoid(_dot_)com> 
<xsl-list-service(_at_)lists(_dot_)mulberrytech(_dot_)com 
<mailto:xsl-list-service(_at_)lists(_dot_)mulberrytech(_dot_)com>> wrote:

Hi
 
An application I am working on contains a large number of source documents 
which are all run through the same series of transformations. While 
initially the build process didn't take long the cost of repeatedly 
initialising the XSL processor soon adds up, so I am looking at ways to 
streamline it.
 
Our processor of choice is Saxon (currently we are using 8.7.3) so I can 
shift this question to the Saxon list if there are extensions there that are 
relevant.
 
So the question; given a script that essentially includes the following:
 
cd documents
for d in `cat dlist`; do
  cd $d
  for f in `cat flist`; do
    java -jar $SAXONDIR/saxon8.jar  -o  $f.new.xml  $f.xml  
$SCRIPTDIR/transform.xsl  doc=$d  file=$f
  done
done
 
is there a mechanism which would allow a single Java process to perform the 
equivalent?
 
Thanks
T
 
XSL-List info and archive <http://www.mulberrytech.com/xsl/xsl-list>
EasyUnsubscribe <http://lists.mulberrytech.com/unsub/xsl-list/293509> (by 
email <>)

XSL-List info and archive <http://www.mulberrytech.com/xsl/xsl-list>
EasyUnsubscribe <http://lists.mulberrytech.com/unsub/xsl-list/782854> (by 
email <applewebdata://ECBB98FE-8171-49BF-A36D-41F19484D8D7>)


-- 
Cheers,
Dimitre Novatchev
---------------------------------------
Truly great madness cannot be achieved without significant intelligence.
---------------------------------------
To invent, you need a good imagination and a pile of junk
-------------------------------------
Never fight an inanimate object
-------------------------------------
To avoid situations in which you might make mistakes may be the
biggest mistake of all
------------------------------------
Quality means doing it right when no one is looking.
-------------------------------------
You've achieved success in your field when you don't know whether what you're 
doing is work or play
-------------------------------------
To achieve the impossible dream, try going to sleep.
-------------------------------------
Facts do not cease to exist because they are ignored.
-------------------------------------
Typing monkeys will write all Shakespeare's works in 200yrs.Will they write 
all patents, too? :)
-------------------------------------
Sanity is madness put to good use.
-------------------------------------
I finally figured out the only reason to be alive is to enjoy it.
 
XSL-List info and archive <http://www.mulberrytech.com/xsl/xsl-list>
EasyUnsubscribe <http://lists.mulberrytech.com/unsub/xsl-list/293509> (by 
email <>)
--~----------------------------------------------------------------
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
EasyUnsubscribe: http://lists.mulberrytech.com/unsub/xsl-list/1167547
or by email: xsl-list-unsub(_at_)lists(_dot_)mulberrytech(_dot_)com
--~--
<Prev in Thread] Current Thread [Next in Thread>