Postgres Pro
Downloads
Home   >   mailing lists

Thread: Re: doc phrase: "inheritance child"

Re: doc phrase: "inheritance child"

From
Ashutosh Bapat
Date:
Hi Justin,
 
@@ -7306,7 +7306,7 @@ SCRAM-SHA-256$<replaceable>&lt;iteration count&gt;</replaceable>:<replaceable>&l
   <para>
    Normally there is one entry, with <structfield>stainherit</structfield> =
    <literal>false</literal>, for each table column that has been analyzed.
-   If the table has inheritance children, a second entry with
+   If the table has inheritance children or partitions, a second entry with
    <structfield>stainherit</structfield> = <literal>true</literal> is also created.  This row
    represents the column's statistics over the inheritance tree, i.e.,
    statistics for the data you'd see with

For partitioned tables only the second entry makes sense. IIRC, we had done some work to remove the first entry. Can you please check whether a partitioned table also has two entries?

       <para>
-       If true, the stats include inheritance child columns, not just the
+       If true, the stats include child tables, not just the

We are replacing columns with tables; is that intentional?

Partitioned tables do not have their own stats, it's just aggregated partition stats.

-   If the table has inheritance children, a second entry with
+   If the table has inheritance children or partitions, a second entry with
    <structfield>stxdinherit</structfield> = <literal>true</literal> is also created.
    This row represents the statistics object over the inheritance tree, i.e.,

Similar to the first comment. s/inheritance tree/inheritance or partition tree/ ?

 
-       If true, the stats include inheritance child columns, not just the
+       If true, the stats include child childs, not just the
        values in the specified relation
       </para></entry>
      </row>
@@ -13152,7 +13152,7 @@ SELECT * FROM pg_locks pl LEFT JOIN pg_prepared_xacts ppx
        <structfield>inherited</structfield> <type>bool</type>
       </para>
       <para>
-       If true, this row includes inheritance child columns, not just the
+       If true, this row includes child tables, not just the
        values in the specified table
       </para></entry>
      </row>

Replacing inheritance child "column" with "tables", is that intentional?

Are these all the places where child/children need to be replaced by partitions?

Now that the feature is old and also being used widely, it probably makes sense to mention partition where inheritance children is mentioned, if this double mention makes sense. But I think it's more than just the replacement. We need to rewrite or make modified copies of some of the sentences or paragraphs entirely. Esp. the things that apply to inheritance may not be applicable as is to partitioning and vice versa. We may be required to replace inheritance tree with partition tree in the nearby sentences.


--
Best Wishes,
Ashutosh

Re: doc phrase: "inheritance child"

From
Amit Langote
Date:
Hi,

On Wed, May 25, 2022 at 1:30 PM Ashutosh Bapat
<ashutosh.bapat@enterprisedb.com> wrote:
> @@ -7306,7 +7306,7 @@ SCRAM-SHA-256$<replaceable><iteration count></replaceable>:<replaceable>&l
>    <para>
>     Normally there is one entry, with <structfield>stainherit</structfield> =
>     <literal>false</literal>, for each table column that has been analyzed.
> -   If the table has inheritance children, a second entry with
> +   If the table has inheritance children or partitions, a second entry with
>     <structfield>stainherit</structfield> = <literal>true</literal> is also created.  This row
>     represents the column's statistics over the inheritance tree, i.e.,
>     statistics for the data you'd see with
>
> For partitioned tables only the second entry makes sense. IIRC, we had done some work to remove the first entry. Can
youplease check whether a partitioned table also has two entries?
 

Don't think we've made any changes yet that get rid of the parent
partitioned table's entry in pg_statistic:

create table foo (a int) partition by list (a);
create table foo1 partition of foo for values in (1);
analyze foo;
select starelid::regclass, stainherit from pg_statistic where
starelid::regclass in (select relid from pg_partition_tree('foo'));
 starelid | stainherit
----------+------------
 foo      | t
 foo1     | f
(2 rows)

Maybe you're thinking of RangeTblEntry that the planner makes 2 copies
for inheritance parents, but only 1 for partition parents as of
e8d5dd6be79.

--
Thanks, Amit Langote
EDB: http://www.enterprisedb.com

Re: doc phrase: "inheritance child"

From
Amit Langote
Date:
On Wed, May 25, 2022 at 1:30 PM Ashutosh Bapat
<ashutosh.bapat@enterprisedb.com> wrote:
>        <para>
> -       If true, the stats include inheritance child columns, not just the
> +       If true, the stats include child tables, not just the
>
> We are replacing columns with tables; is that intentional?
>
> Partitioned tables do not have their own stats, it's just aggregated partition stats.
> ...
> -       If true, the stats include inheritance child columns, not just the
> +       If true, the stats include child childs, not just the
>         values in the specified relation
>        </para></entry>
>       </row>
> @@ -13152,7 +13152,7 @@ SELECT * FROM pg_locks pl LEFT JOIN pg_prepared_xacts ppx
>         <structfield>inherited</structfield> <type>bool</type>
>        </para>
>        <para>
> -       If true, this row includes inheritance child columns, not just the
> +       If true, this row includes child tables, not just the
>         values in the specified table
>        </para></entry>
>       </row>
>
> Replacing inheritance child "column" with "tables", is that intentional?

I was a bit confused by these too, though perhaps the original text is
not as clear as it could be?  Would the following be a good rewrite:

If true, the stats cover the contents not only of the specified table,
but also of its child tables or partitions.  (If the table is
partitioned, which contains no data by itself, the stats only cover
the contents of partitions).

Although, maybe the parenthetical is unnecessary.

-- 
Thanks, Amit Langote
EDB: http://www.enterprisedb.com

Re: doc phrase: "inheritance child"

From
Justin Pryzby
Date:
On Fri, May 27, 2022 at 03:22:38PM +0900, Amit Langote wrote:
> On Wed, May 25, 2022 at 1:30 PM Ashutosh Bapat <ashutosh.bapat@enterprisedb.com> wrote:
> >        <para>
> > -       If true, the stats include inheritance child columns, not just the
> > +       If true, the stats include child tables, not just the
> >
> > We are replacing columns with tables; is that intentional?
> >
> > Partitioned tables do not have their own stats, it's just aggregated partition stats.
> > ...
> > -       If true, the stats include inheritance child columns, not just the
> > +       If true, the stats include child childs, not just the
> >         values in the specified relation
> >        </para></entry>
> >       </row>
> > @@ -13152,7 +13152,7 @@ SELECT * FROM pg_locks pl LEFT JOIN pg_prepared_xacts ppx
> >         <structfield>inherited</structfield> <type>bool</type>
> >        </para>
> >        <para>
> > -       If true, this row includes inheritance child columns, not just the
> > +       If true, this row includes child tables, not just the
> >         values in the specified table
> >        </para></entry>
> >       </row>
> >
> > Replacing inheritance child "column" with "tables", is that intentional?
> 
> I was a bit confused by these too, though perhaps the original text is
> not as clear as it could be?  Would the following be a good rewrite:

I updated the language to say "values from".  Is this better ?

And rebased to include changes to 401f623c7.

BTW nobody complained about my "child child" typo.

-- 
Justin
Attachment

Re: doc phrase: "inheritance child"

From
Amit Langote
Date:
On Thu, Jun 30, 2022 at 6:55 PM Justin Pryzby <pryzby@telsasoft.com> wrote:
> On Fri, May 27, 2022 at 03:22:38PM +0900, Amit Langote wrote:
> > On Wed, May 25, 2022 at 1:30 PM Ashutosh Bapat <ashutosh.bapat@enterprisedb.com> wrote:
> > >        <para>
> > > -       If true, the stats include inheritance child columns, not just the
> > > +       If true, the stats include child tables, not just the
> > >
> > > We are replacing columns with tables; is that intentional?
> > >
> > > Partitioned tables do not have their own stats, it's just aggregated partition stats.
> > > ...
> > > -       If true, the stats include inheritance child columns, not just the
> > > +       If true, the stats include child childs, not just the
> > >         values in the specified relation
> > >        </para></entry>
> > >       </row>
> > > @@ -13152,7 +13152,7 @@ SELECT * FROM pg_locks pl LEFT JOIN pg_prepared_xacts ppx
> > >         <structfield>inherited</structfield> <type>bool</type>
> > >        </para>
> > >        <para>
> > > -       If true, this row includes inheritance child columns, not just the
> > > +       If true, this row includes child tables, not just the
> > >         values in the specified table
> > >        </para></entry>
> > >       </row>
> > >
> > > Replacing inheritance child "column" with "tables", is that intentional?
> >
> > I was a bit confused by these too, though perhaps the original text is
> > not as clear as it could be?  Would the following be a good rewrite:
>
> I updated the language to say "values from".  Is this better ?

Yes.

-- 
Thanks, Amit Langote
EDB: http://www.enterprisedb.com