Re: PoC/WIP: Extended statistics on expressions

On 1/17/21 6:29 AM, Justin Pryzby wrote:
> On Sun, Jan 17, 2021 at 01:23:39AM +0100, Tomas Vondra wrote:
>> diff --git a/doc/src/sgml/ref/create_statistics.sgml b/doc/src/sgml/ref/create_statistics.sgml
>> index 4363be50c3..5b8eb8d248 100644
>> --- a/doc/src/sgml/ref/create_statistics.sgml
>> +++ b/doc/src/sgml/ref/create_statistics.sgml
>> @@ -21,9 +21,13 @@ PostgreSQL documentation
>>
>> <refsynopsisdiv>
>> <synopsis>
>> +CREATE STATISTICS [ IF NOT EXISTS ] <replaceable class="parameter">statistics_name</replaceable>
>> + ON ( <replaceable class="parameter">expression</replaceable> )
>> + FROM <replaceable class="parameter">table_name</replaceable>
>
>> CREATE STATISTICS [ IF NOT EXISTS ] <replaceable class="parameter">statistics_name</replaceable>
>> [ ( <replaceable class="parameter">statistics_kind</replaceable> [, ... ] ) ]
>> - ON <replaceable class="parameter">column_name</replaceable>, <replaceable class="parameter">column_name</replaceable> [, ...]
>> + ON { <replaceable class="parameter">column_name</replaceable> | ( <replaceable class="parameter">expression</replaceable> ) } [, ...]
>> FROM <replaceable class="parameter">table_name</replaceable>
>> </synopsis>
>
> I think this part is wrong, since it's not possible to specify a single column
> in form#2.
>
> If I'm right, the current way is:
>
> - form#1 allows expression statistics of a single expression, and doesn't
> allow specifying "kinds";
>
> - form#2 allows specifying "kinds", but always computes expression statistics,
> and requires multiple columns/expressions.
>
> So it'd need to be column_name|expression, column_name|expression [,...]
>

Strictly speaking you're probably correct - there should be at least two
elements. But I'm somewhat hesitant about making this more complex,
because it'll be harder to understand.

>> @@ -39,6 +43,16 @@ CREATE STATISTICS [ IF NOT EXISTS ] <replaceable class="parameter">statistics_na
>> database and will be owned by the user issuing the command.
>> </para>
>>
>> + <para>
>> + The <command>CREATE STATISTICS</command> command has two basic forms. The
>> + simple variant allows building statistics for a single expression, does
>> + not allow specifying any statistics kinds and provides benefits similar
>> + to an expression index. The full variant allows defining statistics objects
>> + on multiple columns and expressions, and selecting which statistics kinds will
>> + be built. The per-expression statistics are built automatically when there
>> + is at least one expression.
>> + </para>
>
>> + <varlistentry>
>> + <term><replaceable class="parameter">expression</replaceable></term>
>> + <listitem>
>> + <para>
>> + The expression to be covered by the computed statistics. In this case
>> + only a single expression is required, in which case only the expression
>> + statistics kind is allowed. The order of expressions is insignificant.
>
> I think this part is wrong now ?
> I guess there's no user-facing expression "kind" left in the CREATE command.
> I guess "in which case" means "if only one expr is specified".
> "expression" could be either form#1 or form#2.
>
> Maybe it should just say:
>> + An expression to be covered by the computed statistics.
>
> Maybe somewhere else, say:
>> In the second form of the command, the order of expressions is insignificant.
>

Yeah, this is a leftover from when there was "expressions" kind. I'll
reword this a bit.

thanks

--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company