Stata
-----

Stata_ is a powerful general purpose package for data analysis and data management. 
Stata covers a wide range of statistical techniques. 
The package analyzes data quickly by reading all data into memory which may be managed dynamically during a session.

.. _Stata: https://www.stata.com/

Interactive Mode
################

1. SSH onto the *mercury* login node with (optional) X11 forwarding :code:`ssh -X <BoothID>@mercury.chicagobooth.edu`
2. Request an interactive session on a compute node :code:`srun --pty bash --login`
3. Load the module with the desired version of Stata :code:`module load stata/18.0`
4. Start Stata by typing one of the following:

        * :code:`stata-se` for STATA Special Edition.
        * :code:`stata-mp` for STATA Parallel Edition.
        * :code:`xstata-se` for STATA Special Edition with X-Windows
        * :code:`xstata-mp` for STATA Multiprocessing with X-Windows

Batch Mode
##########

..    #. Navigate to the directory with your Stata program that you want to run.
..    #. Run the desired Batch Mode version of Stata:
..        For regular Stata, type :code:`batch-stata <program.do>` at the prompt, where :code:`program.do` is the name of your program file.
..        For Stata Special Edition, type :code:`batch-stataSE <program.do>` at the prompt, where :code:`program.do` is the name of your program file.
..        For Stata Parallel Edition, type :code:`batch-stataMP <program.do>` at the prompt, where :code:`program.do` is the name of your program file.

.. If the program executes without any syntax errors, Stata creates an output file called :code:`<program>.log` in the current directory and a second output file named batch.out in your home directory.

.. If the program contained errors, an error file named :code:`batch.err` will be created in your home directory.

To run Stata in batch mode, generate a submission script similar to the one below, and run by typing :code:`sbatch submit.sh`.

.. code-block:: bash
    :caption: submit.sh

    #!/bin/bash

    #SBATCH --account=phd
    #SBATCH --mem=1G
    #SBATCH --time=0-01:00:00
    #SBATCH --job-name=example_job

    # Load the necessary software modules
    module load stata/18.0
    
    # create a new scratch directory for this job
    scratch_dir="/scratch/${SLURM_JOB_USER}/${SLURM_JOB_ID}"
    mkdir -p $scratch_dir

    # use scratch dir to store tmp files
    export STATATMP=$scratch_dir

    # run script
    dofile='program.do'
    srun stata-se -b do $dofile

    # remove scratch directory when done
    rm -r $scratch_dir