Mathematica 9 is now available
Services & Resources / Wolfram Forums
-----
 /
MathGroup Archive
2005
*January
*February
*March
*April
*May
*June
*July
*August
*September
*October
*November
*December
*Archive Index
*Ask about this page
*Print this page
*Give us feedback
*Sign up for the Wolfram Insider

MathGroup Archive 2005

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Batch

  • To: mathgroup at smc.vnet.net
  • Subject: [mg60449] Re: [mg60364] Batch
  • From: Maria Cristina Dias Tavares <cristina at dsce.fee.unicamp.br>
  • Date: Fri, 16 Sep 2005 03:50:49 -0400 (EDT)
  • References: <20050915043753.D50C7B150@mfep8.connect.com.au>
  • Sender: owner-wri-mathgroup at wolfram.com

Hi all,

    I thank both Dave and Luc very much, I manage to go one step further.

    I have tryed  SelectionEvaluation before, but I could not identify 
the notebook properly. Now after you told me I tryed with NotebookOpen 
and it couldn't find the file ($Failed). I had to include my working 
directory in the preferences, and now it works.

    I prepare a very simple test, just a notebook calling a second nb 
which calls a third one.  It works perfectly, opened and executed each 
notebook without a problem.

    Then I put these callings inside a loop (Do) which would just print 
the loop iteration number.

    Now I come to another problem. All the expressions inside the loop 
are executed and only afterwards the execution of  the second notebook 
is performed. Only after that the execution of notebook 3 is evaluated.

    This is not what I need. I need that the calling of the notebook 
evalutaion is treated  like if it were a routine, the program is 
deviated to the second notebook and after evaluating it completely it 
would come back to the following command inside the loop.

    Can anyone give me another advice ?

    Thank you very much.

    Regards,

            Maria Cristina

David Annetts wrote:

>Hi Maria,
>
>	    I will put my question in a different way. 
>	
>	    What I do now is to read the data directories/file names in a
>loop and then I start the calculation. The calculation is performed inside a
>Do, where I generate my tables, save them in the appropriate directories,
>and generate the graphics and save them also in different directories.  That
>works, but this is not a well-organized procedure because I have all the
>calculations inside a very large Do loop, that means that my notebook is
>almost a single large Do cell.  
>	
>	    I think it would be much better if I could have my calculation
>inside another notebook and if I call this notebook inside the Do loop.
>	
>	    My first try was to have a notebook to read the data file and
>directory names and the call a second notebook. In this notebook I have part
>of my calculation, rather documented and it called the last  notebook which
>would finish the calculation and return to the next data. In the front end I
>manage to call the second notebook but it was only opened, not evaluated, I
>have to evaluate it manually. Perhaps the problem does not involve batch,
>but rather one notebook calling and evaluating another notebook.
>	
>=======================
>
>Luc's suggestion of SelectionEvaluate[] is a good one, especially if you're
>going to automatically generate the batch process.
>
>Sorry for being so blunt, but it sounds to me like the problem is not so
>much the number of points you have to deal with, nor the fact that data live
>in different directories, but that the the problem of processing has not
>been thought through properly.  You may well be over complicating things.  
>
>Another comment I'd make is to caution against splitting your processing
>over several notebooks.  To me, this is a recipe for disaster since I lose
>the one good feature of notebooks -- to perform processing or calculations
>in one place.  If you must split processing amongst files, consigning
>operations to packages is much prefered.
>
>In light of your question's rephrasing, I'd suggest something different.
>Why not reorganise your notebook using sections along the lines of 
>
>1. initialise (get any packages that are needed & set any constants)
>2. get the data from files
>3. process the data
>4. tabulation
>5. processing 
>6. clean things up
>
>Sections can be expanded or collapsed, depending on the detail you require.
>I have had success with this method, reading multi-channel data from
>hundreds of files in different directories, processing them, then plotting
>them & dumping BMP files for later animations.  It works, even on machines
>with 128Mb RAM.
>
>Debugging is easy since specific errors are easier to find in the notebook.
>Running the notebook, because everything is in the one place, is a matter of
>"evaluate notebook", then letting things run until they are complete, rather
>than waiting until one notebook loads, then evaluating that.
>
>Regards,
>
>Dave.
>  
>


-- 
Profa Maria Cristina Tavares
Faculdade de Engenharia Elétrica e de Computação
UNICAMP / FEEC / DSCE
CP 6101 - CEP 13083-970
tel : (19) 3788 3738
fax : (19) 3289 1395
http://www.dsce.fee.unicamp.br/~cristina


  • Prev by Date: Re: Batch
  • Next by Date: Re: solve for a squared variable
  • Previous by thread: Re: Batch
  • Next by thread: Re: Batch