On Fri, Jun 01, 2018 at 03:00:10PM -0400, Alvaro Herrera wrote:
> On 2018-May-23, Justin Pryzby wrote:
>
> > There's two other, wider changes to consider:
...
>
> > - should we find a unified term for "inheritence-based partitioning" and avoid
> > using the word "partitioning" in that context? For example: "Partitioning
> > can be implemented using table inheritance[...]". One possible phrase
> > currently begin used is: "legacy inheritance method".
>
> Yeah, maybe it'd be a good time to do that. In particular I wondered
> whether the section title "Partitioning and Constraint Exclusion" should
> be changed somehow to note the fact that it's mostly for the legacy
> method.
I made changes to avoid "partition" (which I think should mean a child of
relkind='p', and itself of relkind='r') and "partitioned" (meaning relkind='p'
itself) but left alone most instances of "partitioning".
There's two issues. One is the unfortunately-named, recommended setting of
constraint_exclusion='partition' :(
And one is this, which I think should be disambiguated from native
list/range/hash partition bounds:
Use simple equality conditions for list partitioning, or simple range
tests for range partitioning, as illustrated in the preceding examples.
I'm short on words so maybe someone else can recommend language.
On Fri, Jun 01, 2018 at 02:57:22PM -0400, Alvaro Herrera wrote:
> Pushed. I made a couple of minor changes, in particular I added the
It looks like you also fixed a double negative - thanks.
Justin