Split crosswalk.csv into crosswalks/*.csv

[progval]

Creates one table per platform, and add a script (scripts/aggregate.py) to generate crosswalk.csv.

I wrote scripts/aggregate.py so that its output is exactly the content of crosswalk.csv before this PR; hence the hard-coded OLD_ORDER variable.
I can remove this if you wish.

This is a proposal to solve #204. Feel free to comment on it!

[cboettig]

Yay, nice work! 👏 🍾

Can you also add some information to README.md documenting this? Maybe also a CONTRIBUTING.md issue template noting this structure, to remind anyone filing a PR how to do it as a separate file?

I think this will make the process much easier. We have a few stale PRs and other issues that have been caused by all the collisions of using one master file.

Thoughs @mbjones ?

[progval]

Yay, nice work!

Thanks!

Can you also add some information to README.md documenting this? Maybe also a CONTRIBUTING.md issue template noting this structure, to remind anyone filing a PR how to do it as a separate file?

Done: 65cf67b

[progval]

I just had a thought: when a mapping in crosswalks/ does not provide a term matching a CodeMeta property, should the row still be present?

On the one hand, it makes it obvious in each mapping which properties are missing, but on the other hand it makes many useless lines.

[cboettig]

probably best for the row to be present

but hopefully the table-join script is robust enough (e.g do you use a join on schema term as the key column?) to work either way?

[progval]

but hopefully the table-join script is robust enough (e.g do you use a join on schema term as the key column?) to work either way?

No; it checks all rows are present, to avoid mistakes. I can change it to make a proper join, though.

[cboettig]

Hmm, maybe your way is actually better to be more strict in keeping alll rows.

On Thu, Nov 15, 2018 at 5:28 AM Valentin Lorentz ***@***.***> wrote: but hopefully the table-join script is robust enough (e.g do you use a join on schema term as the key column?) to work either way? No; it checks all rows are present, to avoid mistakes. I can change it to make a proper join, though. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#205 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AANleqK-OXlK6T0zkxf0sasZLCLirGHrks5uvWv3gaJpZM4YZwyW> .

-- --- Carl Boettiger http://carlboettiger.info/

[cboettig]

I think this is a great step forward, so I've gone and merged this in now. Thanks @progval !