[csvddf] tweaks to structure #99

rufuspollock opened this Issue Feb 9, 2014 · 2 comments


None yet

2 participants


Propose these changes:

  • Eliminate nesting of values under dialect key
  • Make version optional

New structure would look like:

  // csvddfVersion is now optional
  "csvddfVersion": 1.0,
  "delimiter": ",",
  "doubleQuote": true,
  "lineTerminator": "\r\n",
  "quoteChar": "\"",
  "skipInitialSpace": true

Client specs (e.g. Tabular Data Package) would probably nest this info under a key named dialect or similar but that would be up to them.

Version change: csvddfVersion would be bumped to 2.0

Potential other changes

Make keys optional and specify defaults for their absence (defaults would be as in the example above).

Why make the change:

  • Unnecessary nesting is annoying
  • Requirement to have csvddfVersion is cumbersome for the many circumstances where you may want to use the keys from this spec but wouldn't be able to be conformant

@ldodds wdyt?

ldodds commented Feb 10, 2014

I agree. For the datapackage validator I did, I assumed that the dialect key in the package referred to the sub key, rather than nesting (I couldn't find examples).

So a flatter structure with suggestion of a default key name (dialect) for embedding sounds good to me.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment