- From: Gregg Kellogg <gregg@greggkellogg.net>
- Date: Sun, 6 Apr 2014 15:24:05 -0700
- To: David Booth <david@dbooth.org>
- Cc: "public-csv-wg@w3.org" <public-csv-wg@w3.org>, Jeni Tennison <jeni@jenitennison.com>, Gregg Kellogg <gregg@greggkellogg.com>
On Apr 6, 2014, at 2:54 PM, David Booth <david@dbooth.org> wrote:
> Hi Jeni and Gregg,
>
> First of all, very nice work on this!
> http://w3c.github.io/csvw/syntax/
> I haven't read it in detail, but I do have a few comments/suggestions, which I will break into separate messages.
>
> Section 2.1 Core Data Model
> http://w3c.github.io/csvw/syntax/#h3_core-model
> says "Some of these fields may be null fields". What exactly *is* a null field? I think this needs to be defined. Is it possible, in the absence of assistive metadata, for a CSV+ document to have a null field? In a CSV document, how would a null field be distinguishable from a field consisting of an empty string? Or are they the same? They certainly are not the same in a database.
(FYI: The work is almost entirely Jeni's, I just contributed the EBNF).
Not an official response, but I think that any field delimited by separators (",") is empty, and can't be null. If a record had viewer fields than columns, then the missing fields would be null, but this would also be an illegal CSV.
[[[
4.4 Lines
Each line of a CSV+ file must contain the same number of comma-separated values.
]]]
Gregg
> Thanks!
> David
>
>
Received on Sunday, 6 April 2014 22:24:36 UTC