These are chat archives for frictionlessdata/chat

5th
Sep 2018
Zane Selvans
@zaneselvans
Sep 05 2018 16:19
As a new user of these tools, something I'm noticing is the dependence on the ordering of resources in a data package, and fields in a tabular data resource. This feels a little concerning, since the contents of a package or tabular resource don't seem intrinsically ordered, and a script that for processing/annotating them would potentially become invalid if the same data were re-ordered. Given that resources and fields must have name attributes as well, I'm curious if there's a preferred idiom for referring to these list elements based on their associated names, rather than their list order, to avoid this order dependence in code that works with the data package. The examples I'm seeing seem to just use the ordered list index.
Dominik Moritz
@domoritz
Sep 05 2018 21:12
@/all Google released a dataset search and the comments on HN are resonating with a lot of what people are thinking about in this group: https://news.ycombinator.com/item?id=17919297.
Zane Selvans
@zaneselvans
Sep 05 2018 23:50
Google is one of the backers of Frictionless Data, aren't they? Data Packages seem pretty ideal for being indexed by a data search engine.