On 18 January 2014 03:07, Marko Tiikkaja <marko@joh.to> wrote:
> On 1/12/14, 5:53 AM, I wrote:
>>
>> On 1/9/14, 2:57 PM, Dean Rasheed wrote:
>>>
>>> How it should behave for multi-dimensional arrays is less clear, but
>>> I'd argue that it should return the total number of elements, i.e.
>>> cardinality('{{1,2},{3,4}}'::int[][]) = 4. That would make it
>>> consistent with the choices we've already made for unnest() and
>>> ordinality:
>>> - cardinality(foo) = (select count(*) from unnest(foo)).
>>> - unnest with ordinality would always result in ordinals in the range
>>> [1, cardinality].
>>
>>
>> Ignoring my proposal, this seems like the most reasonable option. I'll
>> send an updated patch along these lines.
>
>
> Here's the patch as promised. Thoughts?
>
A couple of points:
The answer for empty (zero dimensional) arrays is wrong --- you need
special case handling for this case to return 0. In fact why not
simply use ArrayGetNItems()?
In the docs, in the table of array functions, I think it would
probably be useful to make the entry for array_length say "see also
cardinality", otherwise people might just stop reading there. I
suspect that in over 90% of cases, cardinality will be the more
appropriate function to use rather than array_length.
Regards,
Dean