arrays in ADQL/TAP (Was: ObsCore update discussion : adding Axes information in Obscore table)

[Patrick Dowler]

note: Posting this back to DAL since it is about ADQL and TAP :-)

We actually have a few array columns in our tap service (initially 
because we have a veiw called SIAv1 which outputs the metadata and that 
has two columns with arraysize="2"). So, you can have such columns and 
users can select them and in the votable they come out as arrays. There 
is no way to use array columns anywhere but the select list (like BLOB).

I kind of cheat and set the datatype in the tap_schema to votable:double 
and size to something when I have an array column, but that is so my 
query -> votable code will work and I don't really expect any one to 
grok what that means. (This is deep in the fuzzy gray area of the TAP 
spec :-) In my defense, I recall checking and this description also 
produces a schema-valid VOSI-tables document... it is just not what 
someone looking in tap_schema would expect to find.

Anyway, I do agree that arrays can be a useful denormalisation technique 
and I wouldn't be against putting some effort into adding support for 
arrays of primitives (booleans and numbers) to ADQL and TAP.

Pat

PS-Haven't tried tap-lint recently but maybe I will try it again now 
that I fixed the unnecessary failures with quoted column names... would 
be interesting to see if it complains about the array columns.


On 27/04/15 01:42 AM, Mark Taylor wrote:
> On Mon, 27 Apr 2015, Markus Demleitner wrote:
>
>> Now, I admit introducing a second table to obscore is a bit drastic,
>> and opposition to that would be well justified. Which brings us back
>> to the options for denormalization considered in the mail referenced
>> above.
>>
>> Arrays (one each for axis type and axis length) would be almost ideal
>> here, but I don't think we can really formally introduce array
>> processing in ADQL.  Can we?
>
> I will just note here that from the output point of view this would
> not be problematic: VOTables are, like FITS, quite happy with
> array-valued columns.  But I'm not making any comment on the
> impact that the introduction of array processing would have on
> the rest of the TAP/ADQL standards and implementation stack.
>
> --
> Mark Taylor   Astronomical Programmer   Physics, Bristol University, UK
> m.b.taylor at bris.ac.uk +44-117-9288776  http://www.star.bris.ac.uk/~mbt/
>

-- 

Patrick Dowler
Canadian Astronomy Data Centre
National Research Council Canada
5071 West Saanich Road
Victoria, BC V9E 2E7

250-363-0044 (office) 250-363-0045 (fax)