Units needed in ADQL-0.7
womullan at skysrv.pha.jhu.edu
Thu Jan 15 10:57:55 PST 2004
I saw this a 3rd way.
Consider that the SkyNode spec tells you units of columns for each node so
you know the units of 10 for any given node.
Why should the node expect anything except the units it specified ?
If you wish to send the same query to 10 nodes then you are either e.g.
in a portal, it is your responsibility to make the unit conversions
Hence everyone may have their own units and units need never be specified in
Now what you are doing below could be done with a region specification and
that has units built in and implies many conversions may need to be done
but even there I have been discussing with Arnold how we might simplify or
limit this for our work in progress..
On Thu, Jan 15, 2004 at 05:05:19PM +0000, Clive Page wrote:
> It seems to me (following valuable discussions with Guy Rixon, Patricio
> Ortiz, Tony Linde and others) that we cannot ignore the need for physical
> units in our queries. For example if one specifies (using SkyQL to avoid
> the verboseness of XML) something like:
> SELECT * FROM sometable t WHERE ABS(t.galactic_latitude) > 10.0 ;
> The meaning to an astronomer is fairly clear: you only want sources more
> than 10 degrees from the galactic plane. But how does the specifier, or
> the database access layer, know that 10.0 means degrees and not say
> radians or pixels? There are only two possibilities that I can think of:
> (a) the physical units for each column must be standardised so that
> everyone in the community knows that galactic latitude must be in
> degrees and any data centre using anything different has a mandatory
> permanent conversion job to do.
> (b) we need a way of specifying units in the string (and in the XML).
> I think that (a) is too difficult, as there are just too many different
> columns and possible units, and argument as to what should be the
> "standard" would never cease. Using (b) we need a notation for units.
> The nearest thing to a standard notation that has so far been proposed is
> in section 3.2 of this document from CDS:
> I think this looks a pretty good solution to the problem.
> Of course we still need a way of marking the syntax of the string in the
> pseudo-sql, for example in brackets, it would become
> SELECT * FROM sometable t WHERE ABS(t.galactic_latitude) > 10.0[deg] ;
> The data access layer at each site would have to interpret the units and
> do the appropriate conversion, so that if its table used radians, it would
> simply provide the conversion on the fly.
> Clive Page
> Dept of Physics & Astronomy,
> University of Leicester, Tel +44 116 252 3551
> Leicester, LE1 7RH, U.K. Fax +44 116 252 3311
More information about the voql