mirrors/postgres
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Doc: improve partitioning discussion in ddl.sgml.

Browse Source 
This started with the intent to explain that range upper bounds
are exclusive, which previously you could only find out by reading
the CREATE TABLE man page.  But I soon found that section 5.11
really could stand a fair amount of editorial attention.  It's
apparently been revised several times without much concern for
overall flow, nor careful copy-editing.

Back-patch to v11, which is as far as the patch goes easily.

Per gripe from Edson Richter.  Thanks to David Johnston for review.

Discussion: https://postgr.es/m/DM6PR13MB3988736CF8F5DC5720440231CFE60@DM6PR13MB3988.namprd13.prod.outlook.com
This commit is contained in:
Tom Lane 2020-11-14 13:09:53 -05:00
parent 788dd0b839
commit dbca94510c
1 changed files with 203 additions and 168 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

371
doc/src/sgml/ddl.sgml
View File

@ -3543,8 +3543,8 @@ VALUES ('Albany', NULL, NULL, 'NY');
<para> <para>
Query performance can be improved dramatically in certain situations, Query performance can be improved dramatically in certain situations,
particularly when most of the heavily accessed rows of the table are in a particularly when most of the heavily accessed rows of the table are in a
single partition or a small number of partitions. The partitioning single partition or a small number of partitions. Partitioning
substitutes for leading columns of indexes, reducing index size and effectively substitutes for the upper tree levels of indexes,
making it more likely that the heavily-used parts of the indexes making it more likely that the heavily-used parts of the indexes
fit in memory. fit in memory.
</para> </para>
@ -3553,18 +3553,20 @@ VALUES ('Albany', NULL, NULL, 'NY');
<listitem> <listitem>
<para> <para>
When queries or updates access a large percentage of a single When queries or updates access a large percentage of a single
partition, performance can be improved by taking advantage partition, performance can be improved by using a
of sequential scan of that partition instead of using an sequential scan of that partition instead of using an
index and random access reads scattered across the whole table. index, which would require random-access reads scattered across the
whole table.
</para> </para>
</listitem> </listitem>
<listitem> <listitem>
<para> <para>
Bulk loads and deletes can be accomplished by adding or removing Bulk loads and deletes can be accomplished by adding or removing
partitions, if that requirement is planned into the partitioning design. partitions, if the usage pattern is accounted for in the
Doing <command>ALTER TABLE DETACH PARTITION</command> or dropping an individual partitioning design. Dropping an individual partition
partition using <command>DROP TABLE</command> is far faster than a bulk using <command>DROP TABLE</command>, or doing <command>ALTER TABLE
DETACH PARTITION</command>, is far faster than a bulk
operation. These commands also entirely avoid the operation. These commands also entirely avoid the
<command>VACUUM</command> overhead caused by a bulk <command>DELETE</command>. <command>VACUUM</command> overhead caused by a bulk <command>DELETE</command>.
</para> </para>
@ -3577,7 +3579,7 @@ VALUES ('Albany', NULL, NULL, 'NY');
</listitem> </listitem>
</itemizedlist> </itemizedlist>
The benefits will normally be worthwhile only when a table would These benefits will normally be worthwhile only when a table would
otherwise be very large. The exact point at which a table will otherwise be very large. The exact point at which a table will
benefit from partitioning depends on the application, although a benefit from partitioning depends on the application, although a
rule of thumb is that the size of the table should exceed the physical rule of thumb is that the size of the table should exceed the physical
@ -3599,6 +3601,13 @@ VALUES ('Albany', NULL, NULL, 'NY');
the ranges of values assigned to different partitions. For the ranges of values assigned to different partitions. For
example, one might partition by date ranges, or by ranges of example, one might partition by date ranges, or by ranges of
identifiers for particular business objects. identifiers for particular business objects.
Each range's bounds are understood as being inclusive at the
lower end and exclusive at the upper end. For example, if one
partition's range is from <literal>1</literal>
to <literal>10</literal>, and the next one's range is
from <literal>10</literal> to <literal>20</literal>, then
value <literal>10</literal> belongs to the second partition not
the first.
</para> </para>
</listitem> </listitem>
</varlistentry> </varlistentry>
@ -3608,7 +3617,7 @@ VALUES ('Albany', NULL, NULL, 'NY');
<listitem> <listitem>
<para> <para>
The table is partitioned by explicitly listing which key values The table is partitioned by explicitly listing which key value(s)
appear in each partition. appear in each partition.
</para> </para>
</listitem> </listitem>
@ -3640,25 +3649,34 @@ VALUES ('Albany', NULL, NULL, 'NY');
<title>Declarative Partitioning</title> <title>Declarative Partitioning</title>
<para> <para>
<productname>PostgreSQL</productname> offers a way to specify how to <productname>PostgreSQL</productname> allows you to declare
divide a table into pieces called partitions. The table that is divided that a table is divided into partitions. The table that is divided
is referred to as a <firstterm>partitioned table</firstterm>. The is referred to as a <firstterm>partitioned table</firstterm>. The
specification consists of the <firstterm>partitioning method</firstterm> declaration includes the <firstterm>partitioning method</firstterm>
and a list of columns or expressions to be used as the as described above, plus a list of columns or expressions to be used
<firstterm>partition key</firstterm>. as the <firstterm>partition key</firstterm>.
</para> </para>
<para> <para>
All rows inserted into a partitioned table will be routed to one of the The partitioned table itself is a <quote>virtual</quote> table having
<firstterm>partitions</firstterm> based on the value of the partition no storage of its own. Instead, the storage belongs
key. Each partition has a subset of the data defined by its to <firstterm>partitions</firstterm>, which are otherwise-ordinary
<firstterm>partition bounds</firstterm>. The currently supported tables associated with the partitioned table.
partitioning methods are range, list, and hash. Each partition stores a subset of the data as defined by its
<firstterm>partition bounds</firstterm>.
All rows inserted into a partitioned table will be routed to the
appropriate one of the partitions based on the values of the partition
key column(s).
Updating the partition key of a row will cause it to be moved into a
different partition if it no longer satisfies the partition bounds
of its original partition.
</para> </para>
<para> <para>
Partitions may themselves be defined as partitioned tables, using what is Partitions may themselves be defined as partitioned tables, resulting
called <firstterm>sub-partitioning</firstterm>. Partitions may have their in <firstterm>sub-partitioning</firstterm>. Although all partitions
must have the same columns as their partitioned parent, partitions may
have their
own indexes, constraints and default values, distinct from those of other own indexes, constraints and default values, distinct from those of other
partitions. See <xref linkend="sql-createtable"/> for more details on partitions. See <xref linkend="sql-createtable"/> for more details on
creating partitioned tables and partitions. creating partitioned tables and partitions.
@ -3666,90 +3684,21 @@ VALUES ('Albany', NULL, NULL, 'NY');
<para> <para>
It is not possible to turn a regular table into a partitioned table or It is not possible to turn a regular table into a partitioned table or
vice versa. However, it is possible to add a regular or partitioned table vice versa. However, it is possible to add an existing regular or
containing data as a partition of a partitioned table, or remove a partitioned table as a partition of a partitioned table, or remove a
partition from a partitioned table turning it into a standalone table; partition from a partitioned table turning it into a standalone table;
see <xref linkend="sql-altertable"/> to learn more about the this can simplify and speed up many maintenance processes.
See <xref linkend="sql-altertable"/> to learn more about the
<command>ATTACH PARTITION</command> and <command>DETACH PARTITION</command> <command>ATTACH PARTITION</command> and <command>DETACH PARTITION</command>
sub-commands. sub-commands.
</para> </para>
<para>
Individual partitions are linked to the partitioned table with inheritance
behind-the-scenes; however, it is not possible to use some of the
generic features of inheritance (discussed below) with declaratively
partitioned tables or their partitions. For example, a partition
cannot have any parents other than the partitioned table it is a
partition of, nor can a regular table inherit from a partitioned table
making the latter its parent. That means partitioned tables and their
partitions do not participate in inheritance with regular tables.
Since a partition hierarchy consisting of the partitioned table and its
partitions is still an inheritance hierarchy, all the normal rules of
inheritance apply as described in <xref linkend="ddl-inherit"/> with
some exceptions, most notably:
<itemizedlist>
<listitem>
<para>
Both <literal>CHECK</literal> and <literal>NOT NULL</literal>
constraints of a partitioned table are always inherited by all its
partitions. <literal>CHECK</literal> constraints that are marked
<literal>NO INHERIT</literal> are not allowed to be created on
partitioned tables.
</para>
</listitem>
<listitem>
<para>
Using <literal>ONLY</literal> to add or drop a constraint on only the
partitioned table is supported as long as there are no partitions. Once
partitions exist, using <literal>ONLY</literal> will result in an error
as adding or dropping constraints on only the partitioned table, when
partitions exist, is not supported. Instead, constraints on the
partitions themselves can be added and (if they are not present in the
parent table) dropped.
</para>
</listitem>
<listitem>
<para>
As a partitioned table does not have any data directly, attempts to use
<command>TRUNCATE</command> <literal>ONLY</literal> on a partitioned
table will always return an error.
</para>
</listitem>
<listitem>
<para>
Partitions cannot have columns that are not present in the parent. It
is not possible to specify columns when creating partitions with
<command>CREATE TABLE</command>, nor is it possible to add columns to
partitions after-the-fact using <command>ALTER TABLE</command>. Tables may be
added as a partition with <command>ALTER TABLE ... ATTACH PARTITION</command>
only if their columns exactly match the parent.
</para>
</listitem>
<listitem>
<para>
You cannot drop the <literal>NOT NULL</literal> constraint on a
partition's column if the constraint is present in the parent table.
</para>
</listitem>
</itemizedlist>
</para>
<para> <para>
Partitions can also be foreign tables, although they have some limitations Partitions can also be foreign tables, although they have some limitations
that normal tables do not; see <xref linkend="sql-createforeigntable"/> for that normal tables do not; see <xref linkend="sql-createforeigntable"/> for
more information. more information.
</para> </para>
<para>
Updating the partition key of a row might cause it to be moved into a
different partition where this row satisfies the partition bounds.
</para>
<sect3 id="ddl-partitioning-declarative-example"> <sect3 id="ddl-partitioning-declarative-example">
<title>Example</title> <title>Example</title>
@ -3770,7 +3719,7 @@ CREATE TABLE measurement (
We know that most queries will access just the last week's, month's or We know that most queries will access just the last week's, month's or
quarter's data, since the main use of this table will be to prepare quarter's data, since the main use of this table will be to prepare
online reports for management. To reduce the amount of old data that online reports for management. To reduce the amount of old data that
needs to be stored, we decide to only keep the most recent 3 years needs to be stored, we decide to keep only the most recent 3 years
worth of data. At the beginning of each month we will remove the oldest worth of data. At the beginning of each month we will remove the oldest
month's data. In this situation we can use partitioning to help us meet month's data. In this situation we can use partitioning to help us meet
all of our different requirements for the measurements table. all of our different requirements for the measurements table.
@ -3782,7 +3731,7 @@ CREATE TABLE measurement (
<orderedlist spacing="compact"> <orderedlist spacing="compact">
<listitem> <listitem>
<para> <para>
Create <structname>measurement</structname> table as a partitioned Create the <structname>measurement</structname> table as a partitioned
table by specifying the <literal>PARTITION BY</literal> clause, which table by specifying the <literal>PARTITION BY</literal> clause, which
includes the partitioning method (<literal>RANGE</literal> in this includes the partitioning method (<literal>RANGE</literal> in this
case) and the list of column(s) to use as the partition key. case) and the list of column(s) to use as the partition key.
@ -3796,30 +3745,15 @@ CREATE TABLE measurement (
) PARTITION BY RANGE (logdate); ) PARTITION BY RANGE (logdate);
</programlisting> </programlisting>
</para> </para>
<para>
You may decide to use multiple columns in the partition key for range
partitioning, if desired. Of course, this will often result in a larger
number of partitions, each of which is individually smaller. On the
other hand, using fewer columns may lead to a coarser-grained
partitioning criteria with smaller number of partitions. A query
accessing the partitioned table will have to scan fewer partitions if
the conditions involve some or all of these columns.
For example, consider a table range partitioned using columns
<structfield>lastname</structfield> and <structfield>firstname</structfield> (in that order)
as the partition key.
</para>
</listitem> </listitem>
<listitem> <listitem>
<para> <para>
Create partitions. Each partition's definition must specify the bounds Create partitions. Each partition's definition must specify bounds
that correspond to the partitioning method and partition key of the that correspond to the partitioning method and partition key of the
parent. Note that specifying bounds such that the new partition's parent. Note that specifying bounds such that the new partition's
values will overlap with those in one or more existing partitions will values would overlap with those in one or more existing partitions will
cause an error. Inserting data into the parent table that does not map cause an error.
to one of the existing partitions will cause an error; an appropriate
partition must be added manually.
</para> </para>
<para> <para>
@ -3830,10 +3764,9 @@ CREATE TABLE measurement (
</para> </para>
<para> <para>
It is not necessary to create table constraints describing partition For our example, each partition should hold one month's worth of
boundary condition for partitions. Instead, partition constraints are data, to match the requirement of deleting one month's data at a
generated implicitly from the partition bound specification whenever time. So the commands might look like:
there is need to refer to them.
<programlisting> <programlisting>
CREATE TABLE measurement_y2006m02 PARTITION OF measurement CREATE TABLE measurement_y2006m02 PARTITION OF measurement
@ -3855,10 +3788,13 @@ CREATE TABLE measurement_y2008m01 PARTITION OF measurement
WITH (parallel_workers = 4) WITH (parallel_workers = 4)
TABLESPACE fasttablespace; TABLESPACE fasttablespace;
</programlisting> </programlisting>
(Recall that adjacent partitions can share a bound value, since
range upper bounds are treated as exclusive bounds.)
</para> </para>
<para> <para>
To implement sub-partitioning, specify the If you wish to implement sub-partitioning, again specify the
<literal>PARTITION BY</literal> clause in the commands used to create <literal>PARTITION BY</literal> clause in the commands used to create
individual partitions, for example: individual partitions, for example:
@ -3870,16 +3806,29 @@ CREATE TABLE measurement_y2006m02 PARTITION OF measurement
After creating partitions of <structname>measurement_y2006m02</structname>, After creating partitions of <structname>measurement_y2006m02</structname>,
any data inserted into <structname>measurement</structname> that is mapped to any data inserted into <structname>measurement</structname> that is mapped to
<structname>measurement_y2006m02</structname> (or data that is directly inserted <structname>measurement_y2006m02</structname> (or data that is
into <structname>measurement_y2006m02</structname>, provided it satisfies its directly inserted into <structname>measurement_y2006m02</structname>,
partition constraint) will be further redirected to one of its which is allowed provided its partition constraint is satisfied)
will be further redirected to one of its
partitions based on the <structfield>peaktemp</structfield> column. The partition partitions based on the <structfield>peaktemp</structfield> column. The partition
key specified may overlap with the parent's partition key, although key specified may overlap with the parent's partition key, although
care should be taken when specifying the bounds of a sub-partition care should be taken when specifying the bounds of a sub-partition
such that the set of data it accepts constitutes a subset of what such that the set of data it accepts constitutes a subset of what
the partition's own bounds allows; the system does not try to check the partition's own bounds allow; the system does not try to check
whether that's really the case. whether that's really the case.
</para> </para>
<para>
Inserting data into the parent table that does not map
to one of the existing partitions will cause an error; an appropriate
partition must be added manually.
</para>
<para>
It is not necessary to manually create table constraints describing
the partition boundary conditions for partitions. Such constraints
will be created automatically.
</para>
</listitem> </listitem>
<listitem> <listitem>
@ -3887,9 +3836,13 @@ CREATE TABLE measurement_y2006m02 PARTITION OF measurement
Create an index on the key column(s), as well as any other indexes you Create an index on the key column(s), as well as any other indexes you
might want, on the partitioned table. (The key index is not strictly might want, on the partitioned table. (The key index is not strictly
necessary, but in most scenarios it is helpful.) necessary, but in most scenarios it is helpful.)
This automatically creates This automatically creates a matching index on each partition, and
one index on each partition, and any partitions you create or attach any partitions you create or attach later will also have such an
later will also contain the index. index.
An index or unique constraint declared on a partitioned table
is <quote>virtual</quote> in the same way that the partitioned table
is: the actual data is in child indexes on the individual partition
tables.
<programlisting> <programlisting>
CREATE INDEX ON measurement (logdate); CREATE INDEX ON measurement (logdate);
@ -3920,7 +3873,7 @@ CREATE INDEX ON measurement (logdate);
<para> <para>
Normally the set of partitions established when initially defining the Normally the set of partitions established when initially defining the
table is not intended to remain static. It is common to want to table is not intended to remain static. It is common to want to
remove old partitions of data and periodically add new partitions for remove partitions holding old data and periodically add new partitions for
new data. One of the most important advantages of partitioning is new data. One of the most important advantages of partitioning is
precisely that it allows this otherwise painful task to be executed precisely that it allows this otherwise painful task to be executed
nearly instantaneously by manipulating the partition structure, rather nearly instantaneously by manipulating the partition structure, rather
@ -3969,8 +3922,10 @@ CREATE TABLE measurement_y2008m02 PARTITION OF measurement
As an alternative, it is sometimes more convenient to create the As an alternative, it is sometimes more convenient to create the
new table outside the partition structure, and make it a proper new table outside the partition structure, and make it a proper
partition later. This allows the data to be loaded, checked, and partition later. This allows new data to be loaded, checked, and
transformed prior to it appearing in the partitioned table: transformed prior to it appearing in the partitioned table.
The <literal>CREATE TABLE ... LIKE</literal> option is helpful
to avoid tediously repeating the parent table's definition:
<programlisting> <programlisting>
CREATE TABLE measurement_y2008m02 CREATE TABLE measurement_y2008m02
@ -3991,27 +3946,29 @@ ALTER TABLE measurement ATTACH PARTITION measurement_y2008m02
<para> <para>
Before running the <command>ATTACH PARTITION</command> command, it is Before running the <command>ATTACH PARTITION</command> command, it is
recommended to create a <literal>CHECK</literal> constraint on the table to recommended to create a <literal>CHECK</literal> constraint on the table to
be attached matching the desired partition constraint. That way, be attached that matches the expected partition constraint, as
the system will be able to skip the scan which is otherwise needed to validate the implicit illustrated above. That way, the system will be able to skip the scan
which is otherwise needed to validate the implicit
partition constraint. Without the <literal>CHECK</literal> constraint, partition constraint. Without the <literal>CHECK</literal> constraint,
the table will be scanned to validate the partition constraint while the table will be scanned to validate the partition constraint while
holding an <literal>ACCESS EXCLUSIVE</literal> lock on that partition holding both an <literal>ACCESS EXCLUSIVE</literal> lock on that partition
and a <literal>SHARE UPDATE EXCLUSIVE</literal> lock on the parent table. and a <literal>SHARE UPDATE EXCLUSIVE</literal> lock on the parent table.
It may be desired to drop the redundant <literal>CHECK</literal> constraint It is recommended to drop the now-redundant <literal>CHECK</literal>
after <command>ATTACH PARTITION</command> is finished. constraint after <command>ATTACH PARTITION</command> is finished.
</para> </para>
<para> <para>
As explained above, it is possible to create indexes on partitioned tables As explained above, it is possible to create indexes on partitioned tables
and they are applied automatically to the entire hierarchy. This is very so that they are applied automatically to the entire hierarchy.
convenient, as not only the existing partitions will become indexed, but This is very
convenient, as not only will the existing partitions become indexed, but
also any partitions that are created in the future will. One limitation is also any partitions that are created in the future will. One limitation is
that it's not possible to use the <literal>CONCURRENTLY</literal> that it's not possible to use the <literal>CONCURRENTLY</literal>
qualifier when creating such a partitioned index. To overcome long lock qualifier when creating such a partitioned index. To avoid long lock
times, it is possible to use <command>CREATE INDEX ON ONLY</command> times, it is possible to use <command>CREATE INDEX ON ONLY</command>
the partitioned table; such an index is marked invalid, and the partitions the partitioned table; such an index is marked invalid, and the partitions
do not get the index applied automatically. The indexes on partitions can do not get the index applied automatically. The indexes on partitions can
be created separately using <literal>CONCURRENTLY</literal>, and later be created individually using <literal>CONCURRENTLY</literal>, and then
<firstterm>attached</firstterm> to the index on the parent using <firstterm>attached</firstterm> to the index on the parent using
<command>ALTER INDEX .. ATTACH PARTITION</command>. Once indexes for all <command>ALTER INDEX .. ATTACH PARTITION</command>. Once indexes for all
partitions are attached to the parent index, the parent index is marked partitions are attached to the parent index, the parent index is marked
@ -4048,18 +4005,22 @@ ALTER INDEX measurement_city_id_logdate_key
<itemizedlist> <itemizedlist>
<listitem> <listitem>
<para> <para>
There is no way to create an Unique constraints (and hence primary keys) on partitioned tables must
exclusion constraint spanning all partitions; it is only possible include all the partition key columns. This limitation exists because
to constrain each leaf partition individually. the individual indexes making up the constraint can only directly
enforce uniqueness within their own partitions; therefore, the
partition structure itself must guarantee that there are not
duplicates in different partitions.
</para> </para>
</listitem> </listitem>
<listitem> <listitem>
<para> <para>
Unique constraints (and hence primary keys) on partitioned tables must There is no way to create an exclusion constraint spanning the
include all the partition key columns. This limitation exists because whole partitioned table. It is only possible to put such a
<productname>PostgreSQL</productname> can only enforce constraint on each leaf partition individually. Again, this
uniqueness in each partition individually. limitation stems from not being able to enforce cross-partition
restrictions.
</para> </para>
</listitem> </listitem>
@ -4081,11 +4042,76 @@ ALTER INDEX measurement_city_id_logdate_key
</listitem> </listitem>
</itemizedlist> </itemizedlist>
</para> </para>
<para>
Individual partitions are linked to their partitioned table using
inheritance behind-the-scenes. However, it is not possible to use
all of the generic features of inheritance with declaratively
partitioned tables or their partitions, as discussed below. Notably,
a partition cannot have any parents other than the partitioned table
it is a partition of, nor can a table inherit from both a partitioned
table and a regular table. That means partitioned tables and their
partitions never share an inheritance hierarchy with regular tables.
</para>
<para>
Since a partition hierarchy consisting of the partitioned table and its
partitions is still an inheritance hierarchy, all the normal rules of
inheritance apply as described in <xref linkend="ddl-inherit"/>, with
a few exceptions:
<itemizedlist>
<listitem>
<para>
Partitions cannot have columns that are not present in the parent. It
is not possible to specify columns when creating partitions with
<command>CREATE TABLE</command>, nor is it possible to add columns to
partitions after-the-fact using <command>ALTER TABLE</command>.
Tables may be added as a partition with <command>ALTER TABLE
... ATTACH PARTITION</command> only if their columns exactly match
the parent.
</para>
</listitem>
<listitem>
<para>
Both <literal>CHECK</literal> and <literal>NOT NULL</literal>
constraints of a partitioned table are always inherited by all its
partitions. <literal>CHECK</literal> constraints that are marked
<literal>NO INHERIT</literal> are not allowed to be created on
partitioned tables.
You cannot drop a <literal>NOT NULL</literal> constraint on a
partition's column if the same constraint is present in the parent
table.
</para>
</listitem>
<listitem>
<para>
Using <literal>ONLY</literal> to add or drop a constraint on only
the partitioned table is supported as long as there are no
partitions. Once partitions exist, using <literal>ONLY</literal>
will result in an error. Instead, constraints on the partitions
themselves can be added and (if they are not present in the parent
table) dropped.
</para>
</listitem>
<listitem>
<para>
As a partitioned table does not have any data itself, attempts to use
<command>TRUNCATE</command> <literal>ONLY</literal> on a partitioned
table will always return an error.
</para>
</listitem>
</itemizedlist>
</para>
</sect3> </sect3>
</sect2> </sect2>
<sect2 id="ddl-partitioning-implementation-inheritance"> <sect2 id="ddl-partitioning-using-inheritance">
<title>Implementation Using Inheritance</title> <title>Partitioning Using Inheritance</title>
<para> <para>
While the built-in declarative partitioning is suitable for most While the built-in declarative partitioning is suitable for most
common use cases, there are some circumstances where a more flexible common use cases, there are some circumstances where a more flexible
@ -4135,8 +4161,8 @@ ALTER INDEX measurement_city_id_logdate_key
<title>Example</title> <title>Example</title>
<para> <para>
We use the non-partitioned <structname>measurement</structname> This example builds a partitioning structure equivalent to the
table above. To implement partitioning using inheritance, use declarative partitioning example above. Use
the following steps: the following steps:
<orderedlist spacing="compact"> <orderedlist spacing="compact">
@ -4148,7 +4174,16 @@ ALTER INDEX measurement_city_id_logdate_key
to be applied equally to all child tables. There is no point in to be applied equally to all child tables. There is no point in
defining any indexes or unique constraints on it, either. For our defining any indexes or unique constraints on it, either. For our
example, the root table is the <structname>measurement</structname> example, the root table is the <structname>measurement</structname>
table as originally defined. table as originally defined:
<programlisting>
CREATE TABLE measurement (
city_id int not null,
logdate date not null,
peaktemp int,
unitsales int
);
</programlisting>
</para> </para>
</listitem> </listitem>
@ -4195,10 +4230,7 @@ CHECK ( outletID BETWEEN 200 AND 300 )
</programlisting> </programlisting>
This is wrong since it is not clear which child table the key This is wrong since it is not clear which child table the key
value 200 belongs in. value 200 belongs in.
</para> Instead, ranges should be defined in this style:
<para>
It would be better to instead create child tables as follows:
<programlisting> <programlisting>
CREATE TABLE measurement_y2006m02 ( CREATE TABLE measurement_y2006m02 (
@ -4271,7 +4303,7 @@ CREATE TRIGGER insert_measurement_trigger
</programlisting> </programlisting>
We must redefine the trigger function each month so that it always We must redefine the trigger function each month so that it always
points to the current child table. The trigger definition does inserts into the current child table. The trigger definition does
not need to be updated, however. not need to be updated, however.
</para> </para>
@ -4736,10 +4768,10 @@ EXPLAIN SELECT count(*) FROM measurement WHERE logdate &gt;= DATE '2008-01-01';
</sect2> </sect2>
<sect2 id="ddl-partitioning-declarative-best-practices"> <sect2 id="ddl-partitioning-declarative-best-practices">
<title>Declarative Partitioning Best Practices</title> <title>Best Practices for Declarative Partitioning</title>
<para> <para>
The choice of how to partition a table should be made carefully as the The choice of how to partition a table should be made carefully, as the
performance of query planning and execution can be negatively affected by performance of query planning and execution can be negatively affected by
poor design. poor design.
</para> </para>
@ -4749,8 +4781,8 @@ EXPLAIN SELECT count(*) FROM measurement WHERE logdate &gt;= DATE '2008-01-01';
by which you partition your data. Often the best choice will be to by which you partition your data. Often the best choice will be to
partition by the column or set of columns which most commonly appear in partition by the column or set of columns which most commonly appear in
<literal>WHERE</literal> clauses of queries being executed on the <literal>WHERE</literal> clauses of queries being executed on the
partitioned table. <literal>WHERE</literal> clause items that match and partitioned table. <literal>WHERE</literal> clauses that are compatible
are compatible with the partition key can be used to prune unneeded with the partition bound constraints can be used to prune unneeded
partitions. However, you may be forced into making other decisions by partitions. However, you may be forced into making other decisions by
requirements for the <literal>PRIMARY KEY</literal> or a requirements for the <literal>PRIMARY KEY</literal> or a
<literal>UNIQUE</literal> constraint. Removal of unwanted data is also a <literal>UNIQUE</literal> constraint. Removal of unwanted data is also a
@ -4767,7 +4799,8 @@ EXPLAIN SELECT count(*) FROM measurement WHERE logdate &gt;= DATE '2008-01-01';
which could result in low cache hit ratios. However, dividing the table which could result in low cache hit ratios. However, dividing the table
into too many partitions can also cause issues. Too many partitions can into too many partitions can also cause issues. Too many partitions can
mean longer query planning times and higher memory consumption during both mean longer query planning times and higher memory consumption during both
query planning and execution. When choosing how to partition your table, query planning and execution, as further described below.
When choosing how to partition your table,
it's also important to consider what changes may occur in the future. For it's also important to consider what changes may occur in the future. For
example, if you choose to have one partition per customer and you example, if you choose to have one partition per customer and you
currently have a small number of large customers, consider the currently have a small number of large customers, consider the
@ -4781,13 +4814,15 @@ EXPLAIN SELECT count(*) FROM measurement WHERE logdate &gt;= DATE '2008-01-01';
<para> <para>
Sub-partitioning can be useful to further divide partitions that are Sub-partitioning can be useful to further divide partitions that are
expected to become larger than other partitions, although excessive expected to become larger than other partitions.
sub-partitioning can easily lead to large numbers of partitions and can Another option is to use range partitioning with multiple columns in
cause the same problems mentioned in the preceding paragraph. the partition key.
Either of these can easily lead to excessive numbers of partitions,
so restraint is advisable.
</para> </para>
<para> <para>
It is also important to consider the overhead of partitioning during It is important to consider the overhead of partitioning during
query planning and execution. The query planner is generally able to query planning and execution. The query planner is generally able to
handle partition hierarchies with up to a few thousand partitions fairly handle partition hierarchies with up to a few thousand partitions fairly
well, provided that typical queries allow the query planner to prune all well, provided that typical queries allow the query planner to prune all
@ -4796,7 +4831,7 @@ EXPLAIN SELECT count(*) FROM measurement WHERE logdate &gt;= DATE '2008-01-01';
performs partition pruning. This is particularly true for the performs partition pruning. This is particularly true for the
<command>UPDATE</command> and <command>DELETE</command> commands. Another <command>UPDATE</command> and <command>DELETE</command> commands. Another
reason to be concerned about having a large number of partitions is that reason to be concerned about having a large number of partitions is that
the server's memory consumption may grow significantly over a period of the server's memory consumption may grow significantly over
time, especially if many sessions touch large numbers of partitions. time, especially if many sessions touch large numbers of partitions.
That's because each partition requires its metadata to be loaded into the That's because each partition requires its metadata to be loaded into the
local memory of each session that touches it. local memory of each session that touches it.
@ -4810,8 +4845,8 @@ EXPLAIN SELECT count(*) FROM measurement WHERE logdate &gt;= DATE '2008-01-01';
either of these two types of workload, it is important to make the right either of these two types of workload, it is important to make the right
decisions early, as re-partitioning large quantities of data can be decisions early, as re-partitioning large quantities of data can be
painfully slow. Simulations of the intended workload are often beneficial painfully slow. Simulations of the intended workload are often beneficial
for optimizing the partitioning strategy. Never assume that more for optimizing the partitioning strategy. Never just assume that more
partitions are better than fewer partitions and vice-versa. partitions are better than fewer partitions, nor vice-versa.
</para> </para>
</sect2> </sect2>