Re: "writable CTEs"

On Tue, Dec 28, 2010 at 03:49:16AM +0000, Peter Geoghegan wrote:
> On 28 December 2010 01:09, Kevin Grittner <Kevin.Grittner@wicourts.gov> wrote:
> > Personally, I think it's worth fixing.  This sort of disjunction
> > between code and documentation can cause confusing for someone
> > trying to get started on hacking.  It is an exception to the
> > otherwise excellent documentation of both the product and the code.
> 
> Hmm. Having looked at the relevant sgml file, queries.sgml, common
> table expressions appear at one point:
> 
>   <indexterm>
>    <primary>common table expression</primary>
>    <see>WITH</see>
>   </indexterm>
> 
> This indicates that the term common table expression should be
> indexed (the dead tree way), which isn't much use for the majority
> of users that access the docs on the web. This term doesn't appear
> in the html source. Perhaps whatever infrastructure we use to render
> the sgml files as html for dot org should produce keyword meta tags
> for indexed terms, in case anyone searches the docs using Altavista.
> More seriously, if we did this I imagine we'd see WITH Queries (for
> example) in the first page of results if we search for "common table
> expression" from dot org directly. The fact that whatever docbook
> tool we use doesn't already do this does suggests that it might not
> be such a good idea. It may not be worth the effort. I've cc'd Thom
> Brown to see what he thinks.
> 
> Attached documentation patch should make things clearer. I haven't
> changed the "queries-with" section to
> "queries-common-table-expression" per David's suggestion for the
> sake of stability. I hesitate to change it without reaching a
> consensus - will this break a lot of links?

I don't see how people can be relying on links to 9.1-to-be's
documentation.

> The main change I've made is: "WITH queries, also referred to as
> Common table expressions or CTEs, provide a way to write subqueries
> for use as part of a larger query". I'm concerned that this might
> not be strictly correct, because the term "WITH query" may not be
> exactly equivalent to the term "CTE" - WITH queries are comprised of
> one or more CTEs, plus a main query.  Or are they?

They are. :)

Cheers,
David.
-- 
David Fetter <david@fetter.org> http://fetter.org/
Phone: +1 415 235 3778  AIM: dfetter666  Yahoo!: dfetter
Skype: davidfetter      XMPP: david.fetter@gmail.com
iCal: webcal://www.tripit.com/feed/ical/people/david74/tripit.ics

Remember to vote!
Consider donating to Postgres: http://www.postgresql.org/about/donate