[OPEN-ILS-DEV] Proposed modification to 1.6.0.4-1.6.1.0-upgrade-db.sql

[Dan Scott]

On 16 November 2010 09:34, Mike Rylander <mrylander at gmail.com> wrote:
> On Mon, Nov 15, 2010 at 6:16 PM, Joe Atzberger <ohiocore at gmail.com> wrote:
>>
>>
>> On Wed, Nov 10, 2010 at 9:40 AM, Mike Rylander <mrylander at gmail.com> wrote:
>>>
>>> On Mon, Nov 8, 2010 at 3:07 PM, Peters, Michael <MRPeters at library.in.gov>
>>> wrote:
>>> > Hi all,
>>> >
>>> > In working through the process of updating a 1.6.0.0 test server to 2.0
>>> > Alpha 5 I ran across a problem with this particular upgrade.  It was
>>> > impossible for me to excute line 531 of the old script:
>>> >
>>> > ALTER TABLE acq.purchase_order ADD COLUMN name TEXT NOT NULL;
>>> >
>>> > I modified line 531 to first give the name column an empty string for a
>>> > name, then later added the “NOT NULL” constraint on this column via the
>>> > new
>>> > line 532.
>>> >
>>>
>>> Because the acq schema is entirely non-authoritative until 2.0, I
>>> suggest we simply delete any data from this table (and others that may
>>> suffer similar issues) during upgrade.
>>>
>>> Thoughts, folks?
>>>
>>
>> Even if the data is "premature", I don't see the point of destroying it when
>> a workaround exists.
>
> At this point we have two (essentially) diametrically opposed opinions
> on the table:
>  * Try to allow data entered into a demo feature to propagate into a
> significantly different (in layout and interpretation) schema -- the
> hope is that the data will be semantically correct and useful.
>  * Wash our hands of demo data in an attempt to reduce pain down the
> road caused by data was created under assumptions that don't hold
> today.
>
> Whatever consensus is arrived at is fine by me, but there will need to
> be a good bit more work put into confirming old data will 1) convert
> and 2) retain proper semantics if we go with the former.  At a glance,
> acq.provider_contact will need similar treatment as it got a NOT NULL
> field without a default as well.  There may be more to do in the
> schema, and I haven't evaluated (and wouldn't be the best to do so)
> the semantic changes.

Are there any actual users of the 1.6.x preview release of
acquisitions? I can't imagine there are, in which case I see no reason
to try to migrate the old schema & any test data that might have been
created. That just seems like more work for no reason.