|
next
|
Subject: XPath Query Editor slow performance... newbie ignorance resolved! Author: Jim Salmons Date: 06 Sep 2007 10:38 AM Originally Posted: 05 Sep 2007 01:32 PM
|
NUTSHELL PROBLEM: Incredibly slow performance of the XPath Query Editor
Document source is product data about tools, attachements, and accessories for a vendor's product line. A Stylus Studio-generated document schema is associated with the file prior to doing the XPath query.
Sample file: 1.4 MB
File lines: 29,837
Document Root: ITEMS
Document being a collection of: ITEM elements
Number of ITEM elements in sample file: 29
Number of ITEM/IMAGES/IMAGE/URL elements in sample file: 4,409
Target XPath Query: /ITEMS/ITEM/IMAGES/URL
Time to execute query: 6:18 min:sec
Time to open result in new XML document: 6:12 min:sec
Actual target document: 44.5 MB
File lines: 854,037
Document Root: ITEMS
Document being a collection of: ITEM elements
Number of ITEM elements in actual file: 3,151
Number of ITEM/IMAGES/IMAGE/URL elements in sample file: 81,573
Target XPath Query: /ITEMS/ITEM/IMAGES/URL
Time to execute query: UNK
Time to open result in new XML document: UNK
Background info on my hardware: XP Pro x32, 3.0+ GB RAM, 1.86 Ghz dual core Intel CPU
QUESTION #1: True/False - Opening an XPath Query Editor query result in a new XML document involves re-executing the query. If so, you might as well skip the Execute Query task and go right to opening the result in a new document to save half the time required to do your work.
If the answer to questions #1 is true, then the UI is unfortunately deceptive. The Execute green arrow and arrow-into-XML_document icon should be grouped in a drop-down icon list so the user can understand that these two tasks are comparable in execution 'load' and only differ in format and location in which the query result is displayed.
QUESTION #2: Will Stylus Studio _ever_ produce a result to my XPath query on the actual 44.5 MB source file (as described above)?
So far, I haven't had the patience/confidence to wait for a result. After well over 1.5 hours of monitoring the animation of the Processing dialog, I pressed Stop to convince myself that the process was not hung. (The CPU load throughout this ordeal was around 50-53% with the Struzzo.exe process using about 256 MB RAM.) The process wasn't hung, but the aborted query result was only a small portion of the final result expected. This makes we fear that, even if Stylus Studio will eventually produce a result, it will be _many_ hours before it completes. How useful is that?
QUESTION #3: I am a novice user of Stylus Studio. What, if anything, am I obviously missing that will optimize my basic use of this powerful tool?
Unless I can get useful levels of performance, I might as well go back to using REBOL to write and run scripts that generate comparable results in a fraction of the time it will take to do the same things in Stylus Studio. I am tending (hoping) that my performance problems are due to my lack of experience with this tool and XML-associated technologies.
Thanks to any and all who have the time and experience to help me get more out of my Stylus Studio use.
--Jim Salmons--
Iowa Tool Inc.
|
|
|
|