Re: ADQL WD v1.05

From: Yuji SHIRASAKI <yuji.shirasaki-at-nao.ac.jp>
Date: Fri, 14 Jul 2006 00:38:26 +0900 (JST)

Hi All,

I thank Noel, Jeff, and Tony for their comments.

I revised the VOQL WD accoring the comments. 

  http://www.ivoa.net/internal/IVOA/IvoaVOQL/ADQL-20060713.doc
  http://www.ivoa.net/internal/IVOA/IvoaVOQL/ADQL-20060713.pdf
Dead line for comments for this document is around July 24.
After that I will again revise the document and prepare for the PR process.

The major changes are summarized here:

  1. Comments MAY be inserted at nay place of the ADQL/s. Comments MAY NOT be conserved between ADQL/s -> ADQL/x translation. Do we need comments in ADQL/x ?
  2. AS is not optional when specifing the alias name. This is to make it clear that which is an alias name.
  3. Some exmamples of mapping ADQL/s to ADQL/x are added in a Appendix C.
  4. Metadata query is still there. I am not convinced that why metadata content must be in interface document rather than query language doc. Interface doc is a place just to define how to access the data resource and how to return the result. On the other hand Query Language needs metadata or schema information of the data set. So I still think that it is natural to define the table metadata content at the VOQL WD.
  5. Some cautions were made to INTO VOSPACE and UPLOAD syntax. Do we need to reconsider ?
  6. Number of table to be specified in Core is ONE.
  7. Enumeration in schema. under consideration.
  8. Access to all the metadata at one call is supported at SkyNode interface.
  9. Syntax of table name qualified by a service identifier is changed.
  10. ADQL schema are available at: http://www.ivoa.net/internal/IVOA/IvoaVOQL/ADQ-Core-v0.2.xsd http://www.ivoa.net/internal/IVOA/IvoaVOQL/ADQ-v1.05.xsd

Yuji Shirasaki.

From: Yuji SHIRASAKI <yuji.shirasaki-at-nao.ac.jp> Subject: ADQL WD v1.05
Date: Sun, 02 Jul 2006 00:41:30 +0900 (JST)

>
> Hi VOQLers,
>
> July is a due date to finalize the VOQL WDs and move to a
> PR process. That is promised at the victoria meeting...
>
> So now I will ask you to review and commnet to the most recent
> version of ADQL WD that is upload at the VOQL twiki page:
>
> http://www.ivoa.net/internal/IVOA/IvoaVOQL/ADQL-20060701.doc
> or
> http://www.ivoa.net/internal/IVOA/IvoaVOQL/ADQL-20060701.pdf
>
> The first dead line for comments is July 7 (in any time zone).
> I will update the draft according to the comments recieved by
> that date in a few days. I will distribute it again, then
> the second dead line for comment will be 1 week later from
> the second distribution (July 19 ?). Final doc will be distributed
> around July 24 and approved by WGs and go to the PR (I hope).
>
> The skynode interface doc follows the same procedure but it will
> be one or two weeks behind the ADQL doc.
>
> Changes from the WD presented at the Victoria meeting are:
>
> 1. ADQL syntax explanation is moved to main body from Appendix.
> Now the Appendix has only the BNF definition, no explantion
> there.
>
> 2. I assingned ADQL extension IDs by three characters and they
> are summarized in table 1.
>
> 3. Data types are summarized in table 2, and allowed operations
> for data/time data types are summarized in table 3.
>
> 4. Aggregate functions are summarized in table 4 and functions
> are summarized in table 5. The functions are categorized to
> a basic function and an advanced function. A basic function
> is not mandatory event in a Core spec, but they are recommended
> to support.
>
> 5. Xmatch related functions are renamed to join_chi2 and
> join_distance.
>
> 6. A section of metadata query is added, where some of the metadata
> tables are defined.
>
> 7. unit is removed from doc as agreed in Victoria.
>
> 8. frame name is mandatory in specifing the region as agreed in Vic.
>
> 9. IS NULL predicate is added to Core spec.
>
> 10. #UPLOAD is renamed to UPLOAD.
>
> 11. Reserved keywords are defined, which are not used for column name,
> table name, and so on.
>
> 12. A delimiter of a delimited identifier is changed from [] to "".
>
> 13. Literal format for date/time data type is defined as agreed in Vic.
>
> 14. What else ?
>
> Your comments wellcom.
>
> Thank you.
>
> Yuji Shirasaki.
>
>
Received on 2006-07-13Z17:39:51