Update created & valid_start date for all AVAs

[alisonsnwong]

The created date and valid_start date have had a little mix-up and should be updated.

The created date should be the date when the ava is published on govinfo/ecfr. It is also the date mentioned in the fr_revision_history. The valid_start date should be the effective date the ava boundaries begin and is mentioned in "Establishment of the X Viticultural Area" pdf on govinfo. Before, the data documentation mentioned that only AVAs with revision history should have a valid_start date but it should be present in all AVAs.

[kurtishouser]

Looking at several of the updated files for PR #944 it appears there is some additional confusion about the dates.

My current understanding:
created: This is the publication date of the document that established the AVA. It should be the same on all current and historical boundaries for an AVA. Correct?

valid_start: This is the effective date of each document as it relates to a particular AVA. It's typically 30 days after the publication date. It is now required for all current and historical boundaries.

valid_end: This is only used for historical boundaries. It should be the day before the next revision. In almost all cases it's currently the effective date of the next revision (valid_start). So these should be corrected as well, yes?

[MicheleTobias]

Thanks, @kurtishouser! You're right about the created column and we're switching that back. In the discussion about that, my team realized that we are trying to capture two things, so we need to think about adding another column (issue forthcoming).

We all agree about the valid_start column.

Do you know if the end date for an AVA boundary that gets revised is officially the day before the new revision takes affect? We had originally decided to put the same date that the new one started to avoid having a gap, but if there's an official way the TTB thinks about it, we should go with that.

[kurtishouser]

@MicheleTobias I can't be certain how the TTB thinks about it but based on the available information I think their focus is on when something starts (the effective date) rather than when something ends; specifically, how it impacts the wine label.

When an AVA name is changed a (d) section is present in the eCFR.

(d) From November 23, 1984, until June 17, 2004, the name of this viticultural area was “Temecula”.
Effective June 18, 2004, this viticulture area is named “Temecula Valley”. Existing certificates of
label approval showing “Temecula” as the appellation of origin will be revoked by operation of
this regulation on June 19, 2006.

However, they're not consistent with the language in this section. Here's another.

(d) From December 21, 1983, until April 23, 2007, the name of this
viticultural area was “Sonoma County Green Valley”. Effective April 23, 2007, this viticulture area
is named “Green Valley of Russian River Valley”. Existing certificates of label approval showing
“Sonoma County Green Valley” as the appellation of origin will be revoked by operation of this
regulation on April 23, 2009.

The first states the day before the effective date of the next amendment. The second states the day of the next amendment.

When there's a change to the boundaries (usually an expansion) there's also a section in the document regarding the Impact on Current Labels.

I was looking at a weekly grocery store circular recently and the effective prices for the two dates given are always explicitly stated as from the first date through the second date so for me using the day before seems more familiar.

Whatever your team decides just make sure the documentation is perfectly clear regarding what the date represents in the valid_end_date so there won't be any confusion later.

My vote is for from...through rather than from...up to.

[kurtishouser]

It is much easier to visually check dates when using from...up to though, especially when looking at everything in the QGIS attribute table.