From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Eric Schulte" Subject: Re: [BABEL] ob-rec.el and some questions Date: Tue, 15 Feb 2011 13:35:00 -0700 Message-ID: <87ei793ukr.fsf@gmail.com> References: <877hd2uxym.fsf@gnu.org> <87sjvp5dzr.fsf@gmail.com> <87zkpxm6zy.fsf@gnu.org> Mime-Version: 1.0 Content-Type: text/plain Return-path: Received: from [140.186.70.92] (port=44252 helo=eggs.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1PpRc6-0002Lf-9J for emacs-orgmode@gnu.org; Tue, 15 Feb 2011 15:35:11 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1PpRc4-0000Gq-TZ for emacs-orgmode@gnu.org; Tue, 15 Feb 2011 15:35:10 -0500 In-Reply-To: <87zkpxm6zy.fsf@gnu.org> (Jose E. Marchesi's message of "Tue, 15 Feb 2011 20:29:21 +0100") List-Id: "General discussions about Org-mode." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: emacs-orgmode-bounces+geo-emacs-orgmode=m.gmane.org@gnu.org Errors-To: emacs-orgmode-bounces+geo-emacs-orgmode=m.gmane.org@gnu.org To: "Jose E. Marchesi" Cc: emacs-orgmode@gnu.org jemarch@gnu.org (Jose E. Marchesi) writes: > > #+begin_src rec :data hackers.rec :fields Name,Email :type Hacker > > Papers = 'requested' && CreatedAt << '01 January 2011' > > #+end_src > > > > Very cool! Now is the time when I ask, would you consider adding this > to the Org-mode core, which requires filling out the FSF copyright > assignment form? http://orgmode.org/worg/org-contribute.html#sec-2 > > Sure. I assigned future changes to Emacs some years ago. > Ah, I should have noticed the @gnu in your email address. That's great, once this file settles down I'll add it to the repository. > > > C-cC-c in that block echoes "Local setup has been refreshed" and > > nothing happens. A workaround would be to use a selection expression > > that always evaluates to "true", such as: > > > > With my attached version of ob-rec.el it is possible to have a > mostly empty body as follows, however, currently > `org-babel-get-src-block-info' *does* require that there be at least > 1 character as well as a newline in the body. This could be > changed, however I'd want to do some local testing before pushing up > such a change, as I fear it may break existing code blocks. > > That looks like a reasonable workaround. > > > - Is there a way to dinamically change the value of the :results > > parameter in the org-babel-execute:rec function? I would like to add > > a :template parameter containing a template for recfmt, and in that > > case the default value "raw" would not be appropriate. > > > > Please see my attached revision of ob-rec.el. I do not believe that the > default value of raw is appropriate currently as Babel should know when > the returned results are a table. I've slightly modified your existing > code so that raw is no longer the default value, and so that tables are > passed back as tables. Does this change subsume your format > question? > > Yes, now it is clear: I have to adapt the returned data to the > result-params settings, and not the other way around. > well put > > Thanks for the fixes :) > > Also, it sounds like there is currently only support for selecting > records *from* a .rec file, do you plan on adding support for inserting > records *into* a .rec file? > > Extracting the records from org tables or raw rec data, you mean? It > would be easy to use csv2rec and recins to implement a recins block: > Yes, I was thinking of something like the following. Rather than create a new block type, simply add a header argument (or body content) that (somehow) indicates that we are writing not reading. So for example writing an Org-mode table into a fresh .rec file would look like the following. #+tblname: org-data | Foo | Bar | |-----+-----| | 1 | a | | 2 | b | | 3 | c | | 4 | d | #+begin_src rec :data something.rec :var data=org-data :write some rec code to dump data into the rec file... #+end_src [...] > > By evaluating all the code blocks in the file in a sequential way, > additional "rec" blocks could then be used to make (and publish) > selections of the table inserted above: > Yes, and once two-way movement of data is possible, then other languages could be used to update the contents of column in rec-files from within an Org-mode document. Running an arbitrary python function over the "Foo" column of a .rec file and saving the results should be as easy as evaluating the write-foos code block below. 1. Select Foo for all records #+source: all-foos #+begin_src rec :data something.rec :fields Foo 1 #+end_src 2. Run all foos through some arbitrary processing in any Babel language #+source: process-foos #+begin_src python :var foos=all-foos do something to foos #+end_src 3. Save the processed values back into the foos column of the original rec file #+source: write-foos #+begin_src rec :data something.rec :fields Foo :write :var foos=process-foos insert foos #+end_src Hope this makes sense. Best -- Eric > > #+begin_src rec :data foo.rec > some selection expression > #+end_src > > #+results > | Title | Author | > | ... | ... | > > > That sounds like fun! org-mode could then be used as a GUI for recutils > :D