home / github / issue_comments

Menu
  • Search all tables
  • GraphQL API

issue_comments: 277759129

This data as json

html_url issue_url id node_id user created_at updated_at author_association body reactions performed_via_github_app issue
https://github.com/pydata/xarray/issues/1249#issuecomment-277759129 https://api.github.com/repos/pydata/xarray/issues/1249 277759129 MDEyOklzc3VlQ29tbWVudDI3Nzc1OTEyOQ== 1217238 2017-02-06T17:52:05Z 2017-02-06T17:52:05Z MEMBER

This was intentional, but is perhaps too magical/overloaded.

Here's the reasoning: 1. If dimension names are not provided for a 1D array, presume that it's sole dimension is the same as it's name. 2. Variables with the same name as a dimension must be coordinates in the xarray data model. Hence make these variables coordinates, even if they appears in the data_vars argument. (This is a bit of a hold from before there was a separate coords argument.)

Note that if you put an array with 2 or more dimensions directly into data_vars/coords you currently get ValueError.

Alternative behaviors that might make more sense: - Raise an error for this behavior, only allowing implicit dimension names when you put an array in coords. - Deprecate implicit dimension names for 1D arrays altogether. Later, switch to labeling arrays with automatic dimension names like the DataArray constructor.

I'm not a huge fan of this second option because if you don't get an error it would be easy to construct datasets with some variables having the wrong dimension names.

{
    "total_count": 0,
    "+1": 0,
    "-1": 0,
    "laugh": 0,
    "hooray": 0,
    "confused": 0,
    "heart": 0,
    "rocket": 0,
    "eyes": 0
}
  205414496
Powered by Datasette · Queries took 0.88ms · About: xarray-datasette