HCSV (Hybrid CSV) – CSV with inline column types and bit-exact floats

(cristian8vasile.substack.com)

1 points | by cristianvasile 6 hours ago ago

2 comments

  • jqpabc123 4 hours ago ago

    Interesting but only moderately so.

    As mentioned in the article, "out-of-band documentation" has traditionally been associated with CSV data transfer. And while this is a valiant effort, the need for this can never be fully eliminated.

    For example, "speed" and "distance" can be expressed and interpreted in a huge variety of different ways. Which distance --- between what points? Is the speed average or maximum?

    In other words, context is still an essential element that remains to be addressed by "out-of-band documentation". While providing this, why not simply include the metadata and tie it to the column names?

    For example, "DistanceA" is shipping distance between the customer and the supplier's wearhouse measured in furlongs and expressed as a fixed 5 digit zero left padded integer.

  • cristianvasile 6 hours ago ago

    [flagged]