Re: Disallow arrays with non-standard lower bounds

On 1/10/14, 4:14 PM, Merlin Moncure wrote:
> On Fri, Jan 10, 2014 at 4:10 PM, Jeff Janes <jeff(dot)janes(at)gmail(dot)com> wrote:
>> On Fri, Jan 10, 2014 at 1:26 PM, Jim Nasby <jim(at)nasby(dot)net> wrote:
>>>
>>> On 1/9/14, 10:58 PM, Tom Lane wrote:
>>>>
>>>> Jim Nasby <jim(at)nasby(dot)net> writes:
>>>>>
>>>>> ISTM that allowing users to pick arbitrary lower array bounds was a huge
>>>>> mistake. I've never seen anyone make use of it, can't think of any
>>>>> legitimate use cases for it, and hate the stupendous amount of extra code
>>>>> needed to deal with it.
>>>>
>>>>
>>>> You lack imagination, sir.
>>>
>>>
>>> Considering what you'd normally want to do in SQL, the only example I can
>>> think of is to not have the argument over 0 vs 1 based.
>>>
>>> Actually, I was thinking there might be some computational problems where
>>> changing lower bound would be nice, but then again, what other languages
>>> actually support this?
>>
>> Perl does, though they regret it bitterly.
>
> What does it matter? Our arrays have had the capability for years and
> years and "because it's cleaner" is simply not justification to break
> people's applications. Why are we even considering this?

Because it's a foot-gun. So far no one has given a legitimate use case for it and supporting it *greatly* complicates iterating over arrays.

Also, just to be clear, I'd be fine with offering a better alternative and leaving existing arrays alone. I don't see any easy way to do that, but maybe someone's got a good idea on that.
--
Jim C. Nasby, Data Architect jim(at)nasby(dot)net
512.569.9461 (cell) http://jim.nasby.net