<?xml version="1.0" encoding="utf-8" ?>
<!DOCTYPE erlref SYSTEM "erlref.dtd">
<erlref>
<header>
<copyright>
<year>2002</year>
<year>2013</year>
<holder>Ericsson AB, All Rights Reserved</holder>
</copyright>
<legalnotice>
The contents of this file are subject to the Erlang Public License,
Version 1.1, (the "License"); you may not use this file except in
compliance with the License. You should have received a copy of the
Erlang Public License along with this software. If not, it can be
retrieved online at http://www.erlang.org/.
Software distributed under the License is distributed on an "AS IS"
basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See
the License for the specific language governing rights and limitations
under the License.
The Initial Developer of the Original Code is Ericsson AB.
</legalnotice>
<title>mnesia_frag_hash</title>
<prepared>Håkan Mattsson</prepared>
<responsible></responsible>
<docno></docno>
<approved></approved>
<checked></checked>
<date>2002-09-10</date>
<rev>A</rev>
<file>mnesia_frag_hash.sgml</file>
</header>
<module>mnesia_frag_hash</module>
<modulesummary>Defines mnesia_frag_hash callback behavior</modulesummary>
<description>
<p>This module defines a callback behavior for user-defined hash
functions of fragmented tables.</p>
<p>Which module that is selected to implement the <c>mnesia_frag_hash</c>
behavior for a particular fragmented table is specified together
with the other <c>frag_properties</c>. The <c>hash_module</c> defines
the module name. The <c>hash_state</c> defines the initial hash state.</p>
<p>This module implements dynamic hashing, which is a kind of hashing
that grows nicely when new fragments are added. It is well
suited for scalable hash tables.</p>
</description>
<funcs>
<func>
<name>init_state(Tab, State) -> NewState | abort(Reason)</name>
<fsummary>Initiates the hash state for a new table.</fsummary>
<type>
<v>Tab = atom()</v>
<v>State = term()</v>
<v>NewState = term()</v>
<v>Reason = term()</v>
</type>
<desc>
<p>Starts when a fragmented table is
created with the function <c>mnesia:create_table/2</c> or
when a normal (unfragmented) table is converted to be a
fragmented table with <c>mnesia:change_table_frag/2</c>.</p>
<p>Notice that the function <c>add_frag/2</c> is started
one time for each of the other fragments (except number 1)
as a part of the table creation procedure.</p>
<p><c>State</c> is the initial value of the <c>hash_state</c>
<c>frag_property</c>. <c>NewState</c> is stored as
<c>hash_state</c> among the other <c>frag_properties</c>.</p>
</desc>
</func>
<func>
<name>add_frag(State) -> {NewState, IterFrags, AdditionalLockFrags} | abort(Reason)</name>
<fsummary>Starts when a new fragment is added to a fragmented table.</fsummary>
<type>
<v>State = term()</v>
<v>NewState = term()</v>
<v>IterFrags = [integer()]</v>
<v>AdditionalLockFrags = [integer()]</v>
<v>Reason = term()</v>
</type>
<desc>
<p>To scale well, it is a good idea to ensure that the
records are evenly distributed over all fragments, including
the new one.</p>
<p><c>NewState</c> is stored as <c>hash_state</c> among the
other <c>frag_properties</c>.</p>
<p>As a part of the <c>add_frag</c> procedure, <c>Mnesia</c> iterates
over all fragments corresponding to the <c>IterFrags</c> numbers
and starts <c>key_to_frag_number(NewState,RecordKey)</c> for
each record. If the new fragment differs from the old
fragment, the record is moved to the new fragment.</p>
<p>As the <c>add_frag</c> procedure is a part of a schema
transaction, <c>Mnesia</c> acquires write locks on the
affected tables. That is, both the fragments corresponding
to <c>IterFrags</c> and those corresponding to
<c>AdditionalLockFrags</c>.</p>
</desc>
</func>
<func>
<name>del_frag(State) -> {NewState, IterFrags, AdditionalLockFrags} | abort(Reason)</name>
<fsummary>Starts when a fragment is deleted from a fragmented table.</fsummary>
<type>
<v>State = term()</v>
<v>NewState = term()</v>
<v>IterFrags = [integer()]</v>
<v>AdditionalLockFrags = [integer()]</v>
<v>Reason = term()</v>
</type>
<desc>
<p><c>NewState</c> is stored as <c>hash_state</c> among the
other <c>frag_properties</c>.</p>
<p>As a part of the <c>del_frag</c> procedure, <c>Mnesia</c> iterates
over all fragments corresponding to the <c>IterFrags</c> numbers
and starts <c>key_to_frag_number(NewState,RecordKey)</c> for
each record. If the new fragment differs from the old
fragment, the record is moved to the new fragment.</p>
<p>Notice that all records in the last fragment must be moved to
another fragment, as the entire fragment is deleted.</p>
<p>As the <c>del_frag</c> procedure is a part of a schema
transaction, <c>Mnesia</c> acquires write locks on the
affected tables. That is, both the fragments corresponding
to <c>IterFrags</c> and those corresponding to
<c>AdditionalLockFrags</c>.</p>
</desc>
</func>
<func>
<name>key_to_frag_number(State, Key) -> FragNum | abort(Reason)</name>
<fsummary>Resolves the key of a record into a fragment number.</fsummary>
<type>
<v>FragNum = integer()()</v>
<v>Reason = term()</v>
</type>
<desc>
<p>Starts whenever <c>Mnesia</c> needs to determine
which fragment a certain record belongs to. It is typically
started at <c>read</c>, <c>write</c>, and <c>delete</c>.</p>
</desc>
</func>
<func>
<name>match_spec_to_frag_numbers(State, MatchSpec) -> FragNums | abort(Reason)</name>
<fsummary>Resolves a <c>MatchSpec</c> into a list of fragment numbers.</fsummary>
<type>
<v>MatcSpec = ets_select_match_spec()</v>
<v>FragNums = [FragNum]</v>
<v>FragNum = integer()</v>
<v>Reason = term()</v>
</type>
<desc>
<p>This function is called whenever <c>Mnesia</c> needs to determine
which fragments that need to be searched for a <c>MatchSpec</c>.
It is typically called by <c>select</c> and
<c>match_object</c>.</p>
</desc>
</func>
</funcs>
<section>
<title>See Also</title>
<p><seealso marker="mnesia:mnesia">mnesia(3)</seealso></p>
</section>
</erlref>
