Re: st: timestamps, GNU/make-style Stata program

[Roger Harbord <roger.harbord@bristol.ac.uk>]

Sounds interesting to me too. As an alternative approach, have you 
considered using the GNU or Un*x "make" command and calling Stata in 
batch-mode to execute do-files from within the makefile? Maybe a make 
command within Stata would be nice but using existing commands is surely 
going to be quicker to develop.. "make" is quite a sophisticated command 
and replicating all its functionality is going to take quite a bit of 
programming i would have thought.

If you're using Windows then there appear to be possibilities such as MinGW 
(Minimalist GNU for Windows) 
<http://sourceforge.net/project/showfiles.php?group_id=2435> which includes 
a "make" command: 
<http://sourceforge.net/project/shownotes.php?release_id=158854>
Haven't used this myself but I used to use Un*x and "make" is something I 
missed when I started using Stata under Windows and you've reminded me of 
that. If I had it i'd certainly use it rather than my present practice of 
writing "master" do-files in which I (try to) document file dependencies 
but have to select and "do" the appropriate bits by hand after I change 
something.

Good luck James - let us know how you get on, whichever way you choose to 
proceed.

Roger.

--On 15 June 2005 10:35 -0400 Jeph Herrin <jherrin@flyingbuttress.net> 
wrote:


> James,
>
> Interesting question. Seems like the easiest way would be
> to capture the output of -ls- and parse the timestamp out
> of that. It is not exceedingly efficient but I think it would
> be platform independent.
>
> cheers,
> Jeph
>
>
> James Muller wrote:
> > Hello List,
> >
> > I'm wishing to do something (described below), but am unsure of exactly
> > how to implement it (question at bottom). Read on.
> >
> > I'm wishing to implement something like GNU/make in a project I'm
> > working on. That is, something that will compare file timestamps to see
> > whether a dataset, for instance, needs to be rebuilt from a set of
> > scripts. If the timestamp on the do files is more recent than that of
> > the dataset it generates then the dataset should be regenerated. This is
> > a very efficient way of resolving dependencies and such.
> >
> > I've been looking hard and can't find anything within Stata that returns
> > the timestamp of a file. The only way I can find to do this on a .dta
> > file is by reading the file contents itself (there is a text-formatted
> > timestamp in the header of the file). I can't find a way to check the
> > timestamp of any other type of file.
> >
> > There are two alternatives I can see:
> > 1. Call the operating system to return the code, try to somehow capture
> > the output.
> > 2. Use the checksum function along with some other files that contain
> > the last known checksum.
> >
> > Assuming there isn't anything within Stata to do this, the operating
> > system method is best, but I'm not sure how to capture the OS output -
> > do something with environment variables, perhaps.
> >
> > So here's what I ask:
> > 1. Has anybody tried to do something similar before? If so, what was
> > your experience?
> > 2. Is there an internal Stata method to read the timestamp of a file?
> > Version 8.2 preferable to version 9.0.
> > 3. If not, does anybody have any experience with capturing operating
> > system output to Stata?
> > 4. Any comments, etc, at all?
> >
> > Cheers in advance
> >
> > James
> >
> > *
> > *   For searches and help try:
> > *   http://www.stata.com/support/faqs/res/findit.html
> > *   http://www.stata.com/support/statalist/faq
> > *   http://www.ats.ucla.edu/stat/stata/
> *
> *   For searches and help try:
> *   http://www.stata.com/support/faqs/res/findit.html
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/