Query Language Proposal

[Kris Zyp]

Within the Persevere community we have been having a lot of discussions
about simplifying and improving the query language for querying JSON
resources. This is what we have put together:

http://www.persvr.org/draft-zyp-rql-00.html

I would put this forward as a query language design possibility, let me
know if you have any thoughts on it.

--
Thanks,
Kris

[Alexandre Morgaut]

could you, please, in ABNF support, replace:

query = and

and = operator *( "&" operator )

by

query = operator *( "&" operator )

thanx ;-)

--
You received this message because you are subscribed to the Google Groups "json-query" group.
To post to this group, send email to json-...@googlegroups.com.
To unsubscribe from this group, send email to json-query+...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/json-query?hl=en.

[Alexandre Morgaut]

As it is meant to be a query language for JSON resources, why are arrays represented as

(toy,food)

instead of 

[toy,food]

http://www.persvr.org/draft-zyp-rql-00.html#anchor5

Sorry if I am brutal, I just tell things as I feel them on first reading.. ;-)

On Jun 29, 2010, at 7:53 PM, Kris Zyp wrote:

[Kris Zyp]

On 6/29/2010 1:52 PM, Alexandre Morgaut wrote:

As it is meant to be a query language for JSON resources, why are arrays represented as

(toy,food)

instead of 

[toy,food]

Because unfortunately all browsers auto-encode [toy,food] -> %5Btoy,food%5D
Consequently it is impossible to consistently get a browser to send a request with "[toy,food]" in it (via XHR or in the address bar), and disambiguate between brackets that are supposed to URL-encoded as part of strings. Therefore the RQL proposal is designed around known acceptable and working URI delimiters.

http://www.persvr.org/draft-zyp-rql-00.html#anchor5

Sorry if I am brutal, I just tell things as I feel them on first reading.. ;-)

Certainly appreciated, love to have your feedback on this.

-- 
Thanks,
Kris

[Dean Landolt]

On Tue, Jun 29, 2010 at 3:52 PM, Alexandre Morgaut <Alexandr...@4d.com> wrote:

As it is meant to be a query language for JSON resources, why are arrays represented as

(toy,food)

instead of 

[toy,food]

http://www.persvr.org/draft-zyp-rql-00.html#anchor5

Sorry if I am brutal, I just tell things as I feel them on first reading.. ;-)

It is intended to be URI-safe. It is for querying JSON resources, but not intended to be JSON itself (also, why not ["toy","food"] if you're going down that path)?

[Alexandre Morgaut]

<value> seems to be  described in the operator section:

"Each argument may be a value (a string of unreserved or URL-encoded characters)"

and there:

http://www.persvr.org/draft-zyp-rql-00.html#anchor29

I think this specification would be clearer if value and typed value  where defined all together so we could check in the TOC where is defined <value>...

[Lakin Wecker]

I'm not sure if this is the right place to announce this, but RQL itself doesn't have a mailing list. We have implemented an MIT licensed python based RQL library:

https://bitbucket.org/lakin.wecker/pyrqlate

Lakin