path: root/system/doc/efficiency_guide/tablesDatabases.xml

<?xml version="1.0" encoding="utf-8" ?>
<!DOCTYPE chapter SYSTEM "chapter.dtd">

<chapter>
  <header>
    <copyright>
      <year>2001</year><year>2016</year>
      <holder>Ericsson AB. All Rights Reserved.</holder>
    </copyright>
    <legalnotice>
      Licensed under the Apache License, Version 2.0 (the "License");
      you may not use this file except in compliance with the License.
      You may obtain a copy of the License at
 
          http://www.apache.org/licenses/LICENSE-2.0

      Unless required by applicable law or agreed to in writing, software
      distributed under the License is distributed on an "AS IS" BASIS,
      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
      See the License for the specific language governing permissions and
      limitations under the License.
    </legalnotice>

    <title>Tables and Databases</title>
    <prepared>Ingela Anderton</prepared>
    <docno></docno>
    <date>2001-08-07</date>
    <rev></rev>
    <file>tablesDatabases.xml</file>
  </header>

  <section>
    <title>Ets, Dets, and Mnesia</title>
    <p>Every example using Ets has a corresponding example in
      Mnesia. In general, all Ets examples also apply to Dets tables.</p>

    <section>
      <title>Select/Match Operations</title>
      <p>Select/match operations on Ets and Mnesia tables can become
        very expensive operations. They usually need to scan the complete
        table. Try to structure the data to minimize the need for select/match
        operations. However, if you require a select/match operation,
	it is still more efficient than using <c>tab2list</c>.
	Examples of this and of how to avoid select/match are provided in
        the following sections. The functions
        <c>ets:select/2</c> and <c>mnesia:select/3</c> are to be preferred
	over <c>ets:match/2</c>, <c>ets:match_object/2</c>, and
      <c>mnesia:match_object/3</c>.</p>
      <p>In some circumstances, the select/match operations do not need
        to scan the complete table.
	For example, if part of the key is bound when searching an
	<c>ordered_set</c> table, or if it is a Mnesia
	table and there is a secondary index on the field that is
	selected/matched. If the key is fully bound, there is
	no point in doing a select/match, unless you have a bag table
	and are only interested in a subset of the elements with
        the specific key.</p>
      <p>When creating a record to be used in a select/match operation, you
        want most of the fields to have the value "_". The easiest and
	fastest way to do that is as follows:</p>
      <pre>
#person{age = 42, _ = '_'}. </pre>
    </section>

    <section>
      <title>Deleting an Element</title>
      <p>The <c>delete</c> operation is considered
        successful if the element was not present in the table. Hence
        all attempts to check that the element is present in the
        Ets/Mnesia table before deletion are unnecessary. Here follows
        an example for Ets tables:</p>
      <p><em>DO</em></p>
      <pre>
...
ets:delete(Tab, Key),
...</pre>
      <p><em>DO NOT</em></p>
      <pre>
...
case ets:lookup(Tab, Key) of
    [] ->
        ok;
    [_|_] ->
        ets:delete(Tab, Key)
end,
...</pre>
    </section>

    <section>
      <title>Fetching Data</title>
      <p>Do not fetch data that you already have.</p>
      <p>Consider that you have a module that handles the abstract data
      type <c>Person</c>. You export the interface function
      <c>print_person/1</c>, which uses the internal functions
      <c>print_name/1</c>, <c>print_age/1</c>, and
      <c>print_occupation/1</c>.</p>
      <note>
        <p>If the function <c>print_name/1</c>, and so on, had been interface
          functions, the situation would have been different, as you
          do not want the user of the interface to know about the
          internal data representation. </p>
      </note>
      <p><em>DO</em></p>
      <code type="erl">
%%% Interface function
print_person(PersonId) ->
    %% Look up the person in the named table person,
    case ets:lookup(person, PersonId) of
        [Person] ->
            print_name(Person),
            print_age(Person),
            print_occupation(Person);
        [] ->
            io:format("No person with ID = ~p~n", [PersonID])
    end.

%%% Internal functions
print_name(Person) -> 
    io:format("No person ~p~n", [Person#person.name]).
                      
print_age(Person) -> 
    io:format("No person ~p~n", [Person#person.age]).

print_occupation(Person) -> 
    io:format("No person ~p~n", [Person#person.occupation]).</code>
      <p><em>DO NOT</em></p>
      <code type="erl">
%%% Interface function
print_person(PersonId) ->
    %% Look up the person in the named table person,
    case ets:lookup(person, PersonId) of
        [Person] ->
            print_name(PersonID),
            print_age(PersonID),
            print_occupation(PersonID);
        [] ->
            io:format("No person with ID = ~p~n", [PersonID])
    end.

%%% Internal functionss
print_name(PersonID) -> 
    [Person] = ets:lookup(person, PersonId),
    io:format("No person ~p~n", [Person#person.name]).

print_age(PersonID) -> 
    [Person] = ets:lookup(person, PersonId),
    io:format("No person ~p~n", [Person#person.age]).

print_occupation(PersonID) -> 
    [Person] = ets:lookup(person, PersonId),
    io:format("No person ~p~n", [Person#person.occupation]).</code>
    </section>

    <section>
      <title>Non-Persistent Database Storage</title>
      <p>For non-persistent database storage, prefer Ets tables over
        Mnesia <c>local_content</c> tables. Even the Mnesia <c>dirty_write</c>
        operations carry a fixed overhead compared to Ets writes.
        Mnesia must check if the table is replicated or has indices,
        this involves at least one Ets lookup for each
        <c>dirty_write</c>. Thus, Ets writes is always faster than
        Mnesia writes.</p>
    </section>

    <section>
      <title>tab2list</title>
      <p>Assuming an Ets table that uses <c>idno</c> as key
        and contains the following:</p>
      <pre>
[#person{idno = 1, name = "Adam",  age = 31, occupation = "mailman"},
 #person{idno = 2, name = "Bryan", age = 31, occupation = "cashier"},
 #person{idno = 3, name = "Bryan", age = 35, occupation = "banker"},
 #person{idno = 4, name = "Carl",  age = 25, occupation = "mailman"}]</pre>
      <p>If you <em>must</em> return all data stored in the Ets table, you
        can use <c>ets:tab2list/1</c>.  However, usually you are only
        interested in a subset of the information in which case
        <c>ets:tab2list/1</c> is expensive. If you only want to extract
        one field from each record, for example, the age of every person,
        then:</p>
      <p><em>DO</em></p>
      <pre>
...
ets:select(Tab,[{ #person{idno='_', 
                          name='_', 
                          age='$1', 
                          occupation = '_'},
                [],
                ['$1']}]),
...</pre>
      <p><em>DO NOT</em></p>
      <pre>
...
TabList = ets:tab2list(Tab),
lists:map(fun(X) -> X#person.age end, TabList),
...</pre>
      <p>If you are only interested in the age of all persons named
        "Bryan", then:</p>
      <p><em>DO</em></p>
      <pre>
...
ets:select(Tab,[{ #person{idno='_', 
                          name="Bryan", 
                          age='$1', 
                          occupation = '_'},
                [],
                ['$1']}]),
...</pre>
      <p><em>DO NOT</em></p>
      <pre>
...
TabList = ets:tab2list(Tab),
lists:foldl(fun(X, Acc) -> case X#person.name of
                                "Bryan" ->
                                    [X#person.age|Acc];
                                 _ ->
                                     Acc
                           end
             end, [], TabList),
...</pre>
      <p><em>REALLY DO NOT</em></p>
      <pre>
...
TabList = ets:tab2list(Tab),
BryanList = lists:filter(fun(X) -> X#person.name == "Bryan" end,
                         TabList),
lists:map(fun(X) -> X#person.age end, BryanList),
...</pre>
      <p>If you need all information stored in the Ets table about
        persons named "Bryan", then:</p>
      <p><em>DO</em></p>
      <pre>
...
ets:select(Tab, [{#person{idno='_', 
                          name="Bryan", 
                          age='_', 
                          occupation = '_'}, [], ['$_']}]),
...</pre>
      <p><em>DO NOT</em></p>
      <pre>
...
TabList = ets:tab2list(Tab),
lists:filter(fun(X) -> X#person.name == "Bryan" end, TabList),
...</pre>
    </section>

    <section>
      <title>Ordered_set Tables</title>
      <p>If the data in the table is to be accessed so that the order
        of the keys in the table is significant, the table type
        <c>ordered_set</c> can be used instead of the more usual
        <c>set</c> table type. An <c>ordered_set</c> is always
        traversed in Erlang term order regarding the key field
        so that the return values from functions such as <c>select</c>,
        <c>match_object</c>, and <c>foldl</c> are ordered by the key
        values. Traversing an <c>ordered_set</c> with the <c>first</c> and
        <c>next</c> operations also returns the keys ordered.</p>
      <note>
        <p>An <c>ordered_set</c> only guarantees that
          objects are processed in <em>key</em> order.
	  Results from functions such as
          <c>ets:select/2</c> appear in <em>key</em> order even if
          the key is not included in the result.</p>
      </note>
    </section>
  </section>

  <section>
    <title>Ets-Specific</title>

    <section>
      <title>Using Keys of Ets Table</title>
      <p>An Ets table is a single-key table (either a hash table or a
        tree ordered by the key) and is to be used as one. In other
        words, use the key to look up things whenever possible. A
        lookup by a known key in a <c>set</c> Ets table is constant and for
        an <c>ordered_set</c> Ets table it is O(logN). A key lookup is always
        preferable to a call where the whole table has to be
        scanned. In the previous examples, the field <c>idno</c> is the
        key of the table and all lookups where only the name is known
        result in a complete scan of the (possibly large) table
        for a matching result.</p>
      <p>A simple solution would be to use the <c>name</c> field as
        the key instead of the <c>idno</c> field, but that would cause
        problems if the names were not unique. A more general solution would
        be to create a second table with <c>name</c> as key and
        <c>idno</c> as data, that is, to index (invert) the table regarding
        the <c>name</c> field. Clearly, the second table would have to be
        kept consistent with the master table. Mnesia can do this
        for you, but a home brew index table can be very efficient
        compared to the overhead involved in using Mnesia.</p>
      <p>An index table for the table in the previous examples would
        have to be a bag (as keys would appear more than once) and can
        have the following contents:</p>
      <pre>
[#index_entry{name="Adam", idno=1},
 #index_entry{name="Bryan", idno=2},
 #index_entry{name="Bryan", idno=3},
 #index_entry{name="Carl", idno=4}]</pre>
      <p>Given this index table, a lookup of the <c>age</c> fields for
        all persons named "Bryan" can be done as follows:</p>
      <pre>
...
MatchingIDs = ets:lookup(IndexTable,"Bryan"),
lists:map(fun(#index_entry{idno = ID}) ->
                 [#person{age = Age}] = ets:lookup(PersonTable, ID),
                 Age
          end,
          MatchingIDs),
...</pre>
      <p>Notice that this code never uses <c>ets:match/2</c> but
        instead uses the <c>ets:lookup/2</c> call. The
        <c>lists:map/2</c> call is only used to traverse the <c>idno</c>s
        matching the name "Bryan" in the table; thus the number of lookups
        in the master table is minimized.</p>
      <p>Keeping an index table introduces some overhead when
        inserting records in the table. The number of operations gained
        from the table must therefore be compared against the number of
        operations inserting objects in the table. However, notice that the
	gain is significant when the key can be used to lookup elements.</p>
    </section>
  </section>

  <section>
    <title>Mnesia-Specific</title>

    <section>
      <title>Secondary Index</title>
      <p>If you frequently do a lookup on a field that is not the
        key of the table, you lose performance using
        "mnesia:select/match_object" as this function traverses the
        whole table. You can create a secondary index instead and
        use "mnesia:index_read" to get faster access, however this
        requires more memory.</p>
	<p><em>Example</em></p>
      <pre>
-record(person, {idno, name, age, occupation}).
        ...
{atomic, ok} = 
mnesia:create_table(person, [{index,[#person.age]},
                              {attributes,
                                    record_info(fields, person)}]),
{atomic, ok} = mnesia:add_table_index(person, age), 
...

PersonsAge42 =
     mnesia:dirty_index_read(person, 42, #person.age),
...</pre>
    </section>

    <section>
      <title>Transactions </title>
      <p>Using transactions is a way to guarantee that the distributed
        Mnesia database remains consistent, even when many different
        processes update it in parallel. However, if you have
        real-time requirements it is recommended to use <c>dirty</c>
        operations instead of transactions. When using <c>dirty</c>
        operations, you lose the consistency guarantee; this is usually
        solved by only letting one process update the table. Other
        processes must send update requests to that process.</p>
	<p><em>Example</em></p>
      <pre>
...
% Using transaction

Fun = fun() ->
          [mnesia:read({Table, Key}),
           mnesia:read({Table2, Key2})]
      end, 

{atomic, [Result1, Result2]}  = mnesia:transaction(Fun),
...

% Same thing using dirty operations
...

Result1 = mnesia:dirty_read({Table, Key}),
Result2 = mnesia:dirty_read({Table2, Key2}),
...</pre>
    </section>
  </section>
</chapter>