Closed cgates closed 9 years ago
I've gotten rid of the dependence on sqlitemagic in the newly-broken-out
repo (https://github.com/swcarpentry/sql-novice-survey) - it tied out
lesson to the Notebook, which R and MATLAB classes don't use, and even
Notebook users have to be pretty advanced to be doing SQL that way.
I've kept the sqlitemagic.py plugin in the new SQL lesson's code repo,
but would be happy to switch to tkf's.
Gotcha. So the new approach is not generating markdown/html from ipython notebooks?
Correct - people are using Notebooks when they actually teach, but using it for the notes as well was excluding a lot of would-be contributors, since diffing and merging IPython Notebooks with Git is... non-trivial.
Understood. Last question and then I will stop molesting you: So when assembling SQL and results with notes, folks are doing that manually?
Yup - although for me, "manually" means running everything inside Emacs and using a macro to copy the SQL over to the other buffer, run it, copy the output, and paste that back.
Upshot: if you're using Anaconda, and trying to run sql sections in IPython, they may not work as expected.
Novice sql lessons IPython notebooks make heavy use of custom Greg Wilson's sqlitemagic which was inspired by (https://github.com/tkf/ipython-sqlitemagic/blob/master/sqlitemagic.py). This is great, but: a) tkf's sqlitemagic is now included in standard Anaconda b) tkf's sqlitemagic has changed his cell/row magics and they don't match Greg's. c) tkf's sitemagic requires install of texttable package (just FYI)
This seems worth fixing, but I wasn't clear on the best way. My seat of the pants workaround was to rename Greg's adaptation to swc_sqlitemagic.py and adjust the notebook to load that. What thinks others? cg