On Mon, Jun 18, 2018 at 6:11 PM, Andres Freund <andres@anarazel.de> wrote:
> On 2018-06-18 18:05:11 -0400, Tom Lane wrote:
>> Andres Freund <andres@anarazel.de> writes:
>> > On 2018-06-18 17:57:04 -0400, Tom Lane wrote:
>> >> I think the actually desirable way to handle this sort of thing is through
>> >> an "index advisor" sort of plugin, which can hide a given index from the
>> >> planner without any globally visible side-effects.
>>
>> > Although I'm a bit doubtful that just shoving this into an extension is
>> > really sufficient. This is an extremely common task.
>>
>> Well, what I was thinking about was that this functionality already
>> exists (I think) in one or more "index advisor" plugins.
>
> They're doing the opposite, right? I.e. they return "hypothetical
> indexes", which then can be used by the planner. None of the ones I've
> seen currently mask out an existing index.
>
>
>> It's possible that they've all bit-rotted for lack of support, which
>> would not speak highly of the demand for the feature.
>
> IDK, the DBA / developer crowd hitting issues like this isn't the same
> as the crowd willing to update an external plugin that doesn't even do
> quite what you want, and was more experimental than anything.
>
Indeed. ISTR a conversation I had with someone on slack earlier this
year about the validity of just manually updating indisvalid as a
means for determining if an index could be safely removed (for the
record, I did not recommend it ;-)
DBA's are often willing to weedwhacker at things in SQL when the
alternative is to learn C.
Robert Treat
http://xzilla.net