Re: st: set max_memory from command line

[Jasper Haller <jasper.haller@gess.uni-mannheim.de>]

Dear Bill,

thank you for your quick response. While I could recreate #4 in your 
e-mail, it only provides an imperfect workaround for my purposes. What I 
ultimately want to do is run a large number of do-files in batch mode, 
and before I begin, I would like to set max_memory once for running all 
these files. If I understand correctly, implementing your workaround #4 
would require me to re-write all these do-files and add the line -set 
max_memory 10g- at the beginning of every do-file. This is precisely 
what I wanted to avoid having to do by choosing the option -permanently-.

Still, your suggestion points towards a solution. In my Python program 
that calls upon my do-files and executes them in batch mode, I can 
probably create a wrapper of sorts that inserts the line -set max_memory 
10g- automatically into a do-file before executing it.

Thanks,
Jasper


On 06.03.2014 16:31, William Gould, StataCorp LP wrote:
> Jasper <jasper.haller@gess.uni-mannheim.de> reports that
> the values of Stata's memory parameters that he has previously
> -set, premanently- are ignored when running Stata for Linux
> in batch mode.
>
> I have recreated the problem.  On my Linux computer,
>
>      1.  I ran Stata interactively and I typed
>
>                . set max_memory 10g, permanently
>
>          I then exited Stata.
>
>
>      2.  I ran Stata interacively again and I typed
>   
>                . query memory
>
>          The 10g max_memory was displayed just as it should be.
>
>          I exited Stata.
>
>
>      3.  I created a file called report.do containing one line, to wit
>
>                ------------------------------ report.do ---
>                query memory
>                ------------------------------ report.do ---
>
>          I then typed at the Unix prompt
>
>               $ stata -b do report
>
>          That ran fine, but when I looked at the resulting report.log,
>          I was disappointed to discover that max_memory was ., not
>          10g.
>
> I was disappointed because that means Stata has a bug and we at StataCorp
> will have to track it down and fix it.
>
> Until we get the problem fixed, I do have a workaround.  The following
> works:
>
>      4.  I created file trythis.do containing
>
>                ------------------------------ trythis.do ---
>                set max_memory 10g
>                query memory
>                ------------------------------ trythis.do ---
>
>          I then typed at the Unix prompt,
>
>               $ stata -b do trythis
>
>          That ran fine as results are just as they should be.
>          -query memory- reported that max_memory was 10g.
>
> I tested #4 using a Stata for Unix with max_memory set permanently to 10g,
> and with max_memory permanently set to 0g.
>
> -- Bill
> wgould@stata.com
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/