<?xml version="1.0" encoding="utf-8"?>
<p>
  <p><var>Database Values</var></p>

  <p>
    Each column in the database accepts a fixed type of data.  The
    currently defined basic types, and their representations, are:
  </p>

  <dl>
    <dt>integer</dt>
    <dd>
      A decimal integer in the range -2**63 to 2**63-1, inclusive.
    </dd>

    <dt>real</dt>
    <dd>
      A floating-point number.
    </dd>

    <dt>Boolean</dt>
    <dd>
      True or false, written <code>true</code> or <code>false</code>, respectively.
    </dd>

    <dt>string</dt>
    <dd>
      An arbitrary Unicode string, except that null bytes are not allowed.
      Quotes are optional for most strings that begin with an English letter
      or underscore and consist only of letters, underscores, hyphens, and
      periods.  However, <code>true</code> and <code>false</code> and strings that match
      the syntax of UUIDs (see below) must be enclosed in double quotes to
      distinguish them from other basic types.  When double quotes are used,
      the syntax is that of strings in JSON, e.g. backslashes may be used to
      escape special characters.  The empty string must be represented as a
      pair of double quotes (<code>&quot;&quot;</code>).
    </dd>

    <dt>UUID</dt>
    <dd>
      Either a universally unique identifier in the style of RFC 4122,
      e.g. <code>f81d4fae-7dec-11d0-a765-00a0c91e6bf6</code>, or an <code>@</code><var>name</var>
      defined by a <code>get</code> or <code>create</code> command within the
      same <code>ovs-vsctl</code> invocation.
    </dd>

  </dl>

  <p>
    Multiple values in a single column may be separated by spaces or a
    single comma.  When multiple values are present, duplicates are not
    allowed, and order is not important.  Conversely, some database
    columns can have an empty set of values, represented as <code>[]</code>, and
    square brackets may optionally enclose other non-empty sets or single
    values as well.
  </p>

  <p>
    A few database columns are ``maps&apos;&apos; of key-value pairs, where the key
    and the value are each some fixed database type.  These are specified
    in the form <var>key</var><code>=</code><var>value</var>, where <var>key</var> and <var>value</var>
    follow the syntax for the column's key type and value type,
    respectively.  When multiple pairs are present (separated by spaces or
    a comma), duplicate keys are not allowed, and again the order is not
    important.  Duplicate values are allowed.  An empty map is represented
    as <code>{}</code>.  Curly braces may optionally enclose non-empty maps as
    well (but use quotes to prevent the shell from expanding
    <code>other-config={0=x,1=y}</code> into <code>other-config=0=x
    other-config=1=y</code>, which may not have the desired effect).
  </p>

  <p><var>Database Command Syntax</var></p>

  <dl>
    <dt>[<code>--if-exists</code>] [<code>--columns=</code><var>column</var>[<code>,</code><var>column</var>]...] <code>list</code> <var>table</var> [<var>record</var>]...</dt>
    <dd>
      <p>
        Lists the data in each specified <var>record</var>.  If no
        records are specified, lists all the records in <var>table</var>.
      </p>
      <p>
        If <code>--columns</code> is specified, only the requested columns are
        listed, in the specified order.  Otherwise, all columns are listed, in
        alphabetical order by column name.
      </p>
      <p>
        Without <code>--if-exists</code>, it is an error if any specified
        <var>record</var> does not exist.  With <code>--if-exists</code>, the command
        ignores any <var>record</var> that does not exist, without producing any
        output.
      </p>
    </dd>

    <dt>[<code>--columns=</code><var>column</var>[<code>,</code><var>column</var>]...] <code>find</code> <var>table</var> [<var>column</var>[<code>:</code><var>key</var>]<code>=</code><var>value</var>]...</dt>
    <dd>
      <p>
        Lists the data in each record in <var>table</var> whose <var>column</var> equals
        <var>value</var> or, if <var>key</var> is specified, whose <var>column</var> contains
        a <var>key</var> with the specified <var>value</var>.  The following operators
        may be used where <code>=</code> is written in the syntax summary:
      </p>
      <dl>
        <dt><code>= != &lt; &gt; &lt;= &gt;=</code></dt>
        <dd>
          <p>
          Selects records in which <var>column</var>[<code>:</code><var>key</var>] equals, does not
          equal, is less than, is greater than, is less than or equal to, or is
          greater than or equal to <var>value</var>, respectively.</p>
          <p>Consider <var>column</var>[<code>:</code><var>key</var>] and <var>value</var> as sets of
          elements.  Identical sets are considered equal.  Otherwise, if the
          sets have different numbers of elements, then the set with more
          elements is considered to be larger.  Otherwise, consider a element
          from each set pairwise, in increasing order within each set.  The
          first pair that differs determines the result.  (For a column that
          contains key-value pairs, first all the keys are compared, and values
          are considered only if the two sets contain identical keys.)
          </p>
        </dd>

        <dt><code>{=} {!=}</code></dt>
        <dd>
          Test for set equality or inequality, respectively.
        </dd>

        <dt><code>{&lt;=}</code></dt>
        <dd>
          Selects records in which <var>column</var>[<code>:</code><var>key</var>] is a subset of
          <var>value</var>.  For example, <code>flood-vlans{&lt;=}1,2</code> selects records in
          which the <code>flood-vlans</code> column is the empty set or contains 1 or 2
          or both.
        </dd>

        <dt><code>{&lt;}</code></dt>
        <dd>
          Selects records in which <var>column</var>[<code>:</code><var>key</var>] is a proper
          subset of <var>value</var>.  For example, <code>flood-vlans{&lt;}1,2</code> selects
          records in which the <code>flood-vlans</code> column is the empty set or
          contains 1 or 2 but not both.
        </dd>

        <dt><code>{&gt;=} {&gt;}</code></dt>
        <dd>
          Same as <code>{&lt;=}</code> and <code>{&lt;}</code>, respectively, except that the
          relationship is reversed.  For example, <code>flood-vlans{&gt;=}1,2</code>
          selects records in which the <code>flood-vlans</code> column contains both 1
          and 2.
        </dd>
      </dl>

      <p>
        The following operators are available only in Open vSwitch 2.16 and
        later:
      </p>

      <dl>
        <dt><code>{in}</code></dt>
        <dd>
          Selects records in which every element in
          <var>column</var>[<code>:</code><var>key</var>] is also in
          <var>value</var>.  (This is the same as <code>{&lt;=}</code>.)
        </dd>

        <dt><code>{not-in}</code></dt>
        <dd>
          Selects records in which every element in
          <var>column</var>[<code>:</code><var>key</var>] is not in
          <var>value</var>.
        </dd>
      </dl>

      <p>
        For arithmetic operators (<code>= != &lt; &gt; &lt;= &gt;=</code>), when <var>key</var> is
        specified but a particular record's <var>column</var> does not contain
        <var>key</var>, the record is always omitted from the results.  Thus, the
        condition <code>other-config:mtu!=1500</code> matches records that have a
        <code>mtu</code> key whose value is not 1500, but not those that lack an
        <code>mtu</code> key.
      </p>

      <p>
        For the set operators, when <var>key</var> is specified but a particular
        record's <var>column</var> does not contain <var>key</var>, the comparison is
        done against an empty set.  Thus, the condition
        <code>other-config:mtu{!=}1500</code> matches records that have a <code>mtu</code>
        key whose value is not 1500 and those that lack an <code>mtu</code> key.
      </p>

      <p>
        Don't forget to escape <code>&lt;</code> or <code>></code> from interpretation by the
        shell.
      </p>

      <p>
        If <code>--columns</code> is specified, only the requested columns are
        listed, in the specified order.  Otherwise all columns are listed, in
        alphabetical order by column name.
      </p>

      <p>
        The UUIDs shown for rows created in the same <code>ovs-vsctl</code>
        invocation will be wrong.
      </p>

    </dd>

    <dt>[<code>--if-exists</code>] [<code>--id=@</code><var>name</var>] <code>get</code> <var>table record</var> [<var>column</var>[<code>:</code><var>key</var>]]...</dt>
    <dd>
      <p>
        Prints the value of each specified <var>column</var> in the given
        <var>record</var> in <var>table</var>.  For map columns, a <var>key</var> may
        optionally be specified, in which case the value associated with
        <var>key</var> in the column is printed, instead of the entire map.
      </p>
      <p>
        Without <code>--if-exists</code>, it is an error if <var>record</var> does not
        exist or <var>key</var> is specified, if <var>key</var> does not exist in
        <var>record</var>.  With <code>--if-exists</code>, a missing <var>record</var>
        yields no output and a missing <var>key</var> prints a blank line.
      </p>
      <p>
        If <code>@</code><var>name</var> is specified, then the UUID for <var>record</var> may be
        referred to by that name later in the same <code>ovs-vsctl</code>
        invocation in contexts where a UUID is expected.
      </p>
      <p>
        Both <code>--id</code> and the <var>column</var> arguments are optional, but
        usually at least one or the other should be specified.  If both are
        omitted, then <code>get</code> has no effect except to verify that
        <var>record</var> exists in <var>table</var>.
      </p>
      <p>
        <code>--id</code> and <code>--if-exists</code> cannot be used together.
      </p>
    </dd>

    <dt>[<code>--if-exists</code>] <code>set</code> <var>table record column</var>[<code>:</code><var>key</var>]<code>=</code><var>value</var>...</dt>
    <dd>
      <p>
        Sets the value of each specified <var>column</var> in the given
        <var>record</var> in <var>table</var> to <var>value</var>.  For map columns, a
        <var>key</var> may optionally be specified, in which case the value
        associated with <var>key</var> in that column is changed (or added, if none
        exists), instead of the entire map.
      </p>
      <p>
        Without <code>--if-exists</code>, it is an error if <var>record</var> does not
        exist.  With <code>--if-exists</code>, this command does nothing if
        <var>record</var> does not exist.
      </p>
    </dd>
    <dt>[<code>--if-exists</code>] <code>add</code> <var>table record column</var> [<var>key</var><code>=</code>]<var>value</var>...</dt>
    <dd>
      <p>
        Adds the specified value or key-value pair to <var>column</var> in
        <var>record</var> in <var>table</var>.  If <var>column</var> is a map, then <var>key</var>
        is required, otherwise it is prohibited.  If <var>key</var> already exists
        in a map column, then the current <var>value</var> is not replaced (use the
        <code>set</code> command to replace an existing value).
      </p>
      <p>
        Without <code>--if-exists</code>, it is an error if <var>record</var> does not
        exist.  With <code>--if-exists</code>, this command does nothing if
        <var>record</var> does not exist.
      </p>
    </dd>

    <dt>
      <p>
        [<code>--if-exists</code>] <code>remove</code> <var>table record column value</var>...
      </p>
      <p>
        [<code>--if-exists</code>] <code>remove</code> <var>table record column key</var>...
      </p>
      <p>
        [<code>--if-exists</code>] <code>remove</code> <var>table record column key</var><code>=</code><var>value</var>...
      </p>
    </dt>
    <dd>
      <p>
        Removes the specified values or key-value pairs from <var>column</var> in
        <var>record</var> in <var>table</var>.  The first form applies to columns that
        are not maps: each specified <var>value</var> is removed from the column.
        The second and third forms apply to map columns: if only a <var>key</var>
        is specified, then any key-value pair with the given <var>key</var> is
        removed, regardless of its value; if a <var>value</var> is given then a
        pair is removed only if both key and value match.
      </p>
      <p>
        It is not an error if the column does not contain the specified key or
        value or pair.
      </p>
      <p>
        Without <code>--if-exists</code>, it is an error if <var>record</var> does not
        exist.  With <code>--if-exists</code>, this command does nothing if
        <var>record</var> does not exist.
      </p>
    </dd>

    <dt>[<code>--if-exists</code>] <code>clear</code> <var>table record column</var>...</dt>
    <dd>
      <p>
        Sets each <var>column</var> in <var>record</var> in <var>table</var> to the empty set
        or empty map, as appropriate.  This command applies only to columns
        that are allowed to be empty.
      </p>
      <p>
        Without <code>--if-exists</code>, it is an error if <var>record</var> does not
        exist.  With <code>--if-exists</code>, this command does nothing if
        <var>record</var> does not exist.
      </p>
    </dd>

    <dt>[<code>--id=(@</code><var>name</var>|<var>uuid</var>)] <code>create</code> <var>table column</var>[<code>:</code><var>key</var>]<code>=</code><var>value</var>...</dt>
    <dd>
      <p>
        Creates a new record in <var>table</var> and sets the initial values of
        each <var>column</var>.  Columns not explicitly set will receive their
        default values.  Outputs the UUID of the new row.
      </p>
      <p>
        If <code>@</code><var>name</var> is specified, then the UUID for the new row may be
        referred to by that name elsewhere in the same <code>\*(PN</code>
        invocation in contexts where a UUID is expected.  Such references may
        precede or follow the <code>create</code> command.
      </p>
      <p>
        If a valid <var>uuid</var> is specified, then it is used as the
        UUID of the new row.
      </p>
      <dl>
        <dt>Caution (ovs-vsctl as example)</dt>
        <dd>
          Records in the Open vSwitch database are significant only when they
          can be reached directly or indirectly from the <code>Open_vSwitch</code>
          table.  Except for records in the <code>QoS</code> or <code>Queue</code> tables,
          records that are not reachable from the <code>Open_vSwitch</code> table are
          automatically deleted from the database.  This deletion happens
          immediately, without waiting for additional <code>ovs-vsctl</code> commands
          or other database activity.  Thus, a <code>create</code> command must
          generally be accompanied by additional commands <var>within the same</var>
          <code>ovs-vsctl</code> <var>invocation</var> to add a chain of references to the
          newly created record from the top-level <code>Open_vSwitch</code> record.
          The <code>EXAMPLES</code> section gives some examples that show how to do
          this.
        </dd>
      </dl>
    </dd>

    <dt>[<code>--if-exists</code>] <code>destroy</code> <var>table record</var>...</dt>
    <dd>
      Deletes each specified <var>record</var> from <var>table</var>.  Unless
      <code>--if-exists</code> is specified, each <var>record</var>s must exist.
    </dd>

    <dt><code>--all destroy</code> <var>table</var></dt>
    <dd>
      <p>
        Deletes all records from the <var>table</var>.
      </p>
      <dl>
        <dt>Caution (ovs-vsctl as example)</dt>
        <dd>
          The <code>destroy</code> command is only useful for records in the <code>QoS</code>
          or <code>Queue</code> tables.  Records in other tables are automatically
          deleted from the database when they become unreachable from the
          <code>Open_vSwitch</code> table.  This means that deleting the last reference
          to a record is sufficient for deleting the record itself.  For records
          in these tables, <code>destroy</code> is silently ignored.  See the
          <code>EXAMPLES</code> section below for more information.
        </dd>
      </dl>
    </dd>

    <dt><code>wait-until</code> <var>table record</var> [<var>column</var>[<code>:</code><var>key</var>]<code>=</code><var>value</var>]...</dt>
    <dd>
      <p>
        Waits until <var>table</var> contains a record named <var>record</var>
        whose <var>column</var> equals <var>value</var> or, if <var>key</var>
        is specified, whose <var>column</var> contains a <var>key</var> with
        the specified <var>value</var>.  This command supports the same
        operators and semantics described for the <code>find</code> command
        above.
      </p>
      <p>
        If no <var>column</var>[<code>:</code><var>key</var>]<code>=</code><var>value</var> arguments are given,
        this command waits only until <var>record</var> exists.  If more than one
        such argument is given, the command waits until all of them are
        satisfied.
      </p>
      <dl>
        <dt>Caution (ovs-vsctl as example)</dt>
        <dd>
          Usually <code>wait-until</code> should be placed at the beginning of a set
          of <code>ovs-vsctl</code> commands.  For example, <code>wait-until bridge br0
          -- get bridge br0 datapath_id</code> waits until a bridge named
          <code>br0</code> is created, then prints its <code>datapath_id</code> column,
          whereas <code>get bridge br0 datapath_id -- wait-until bridge br0</code>
          will abort if no bridge named <code>br0</code> exists when <code>ovs-vsctl</code>
          initially connects to the database.
        </dd>
      </dl>
      <p>
        Consider specifying <code>--timeout=0</code> along with
        <code>--wait-until</code>, to prevent <code>ovs-vsctl</code> from
        terminating after waiting only at most 5 seconds.
      </p>
    </dd>

    <dt><code>comment</code> [<var>arg</var>]...</dt>
    <dd>
    <p>
      This command has no effect on behavior, but any database log record
      created by the command will include the command and its arguments.
    </p>
    </dd>

  </dl>
</p>