I'm using Stata 16 MP GUI with a network license on RHEL. I am connected to that system using MobaXterm on Windows 10. Stata's "memory" of my previous session's preferences--things like font sizes/colors, window sizes, combined layout setting, etc.--seems sporadic and entirely dependent on the specific node (and port?) I'm using. I would like to unify these preferences in a single location, if possible.
I can see that all of the relevant .rc files are stored in my ~/.stata16 directory. What appears to happen is that for every new value of my $DISPLAY environment variable, Stata creates a subdirectory with the value of $DISPLAY as its name, with a stata16.rc file inside. This fragments the memory of my previous sessions' preferences. If I have altered my preferences while using a certain $DISPLAY, Stata will remember those changes. If it is my first time using that value of $DISPLAY, Stata initializes with factory defaults. Note that I do not choose the value of $DISPLAY; the node I use in a given session is assigned by slurm.
I need to go the automatically-saved .rc route because using the "save preferences"/"open preferences" feature does not work as expected in my setup and that seems much more difficult to troubleshoot (possibly an issue with x-forwarding on the system). In contrast, the automatic memory works very well but is fragmented across nodes as described above. I don't think the nodes differ in any way that would make maintaining separate .rc files necessary, as I always use X-forwarding to the same Windows 10 machine and display setup.
My question: Is there a way to instruct Stata to use a certain "master" stata16.rc file I specify, regardless of the particular node Stata is running on? If so, perhaps I could just put that instruction in my profile.do file.
The alternative I see is a little hackish, something like -!mkdir ~/.stata16/\$DISPLAY- and then -!cp ~/.stata16/master_stata16.rc ~/.stata16/\$DISPLAY/stata16.rc- in my profile.do and then restarting Stata if needed. This is less ideal for a number of reasons, including that the preferences would be static.
Related Posts with Unified preference configuration file across nodes on unix system
Percentage explained by the standard deviation of linear combination of independent variablesHello, everyone, I have a question about using standard deviation to explain changes in the dependen…
"codebook, compact" with complete names of variables?Hi all, I use STATA 17 and I do like the "codebook, compact" command to generate a log-file with inf…
Predict residuals after running a FE / RE modelHow can residual centering be done after running a panel data regression? It can be done easily aft…
Inverted U-Shape combined with moderationHello everyone, for the analysis of my data I identified a inverted U-shape as suitable for my mode…
Linear regression with AR(1) disturbanceDear Colleagues, Currently, I am conducting research and selected panel regression (FE) with AR(1) …
Subscribe to:
Post Comments (Atom)
0 Response to Unified preference configuration file across nodes on unix system
Post a Comment