<?xml version="1.0" encoding="utf-8"?>
<!-- $Revision$ -->
<refentry xml:id="datetimeimmutable.createfromformat" xmlns="http://docbook.org/ns/docbook" xmlns:xlink="http://www.w3.org/1999/xlink">
 <refnamediv>
  <refname>DateTimeImmutable::createFromFormat</refname>
  <refname>date_create_immutable_from_format</refname>
  <refpurpose>Parses a time string according to a specified format</refpurpose>
 </refnamediv>

 <refsect1 role="description">
  &reftitle.description;
  <para>&style.oop;</para>
  <methodsynopsis role="DateTimeImmutable">
   <modifier>public</modifier> <modifier>static</modifier> <type class="union"><type>DateTimeImmutable</type><type>false</type></type><methodname>DateTimeImmutable::createFromFormat</methodname>
   <methodparam><type>string</type><parameter>format</parameter></methodparam>
   <methodparam><type>string</type><parameter>datetime</parameter></methodparam>
   <methodparam choice="opt"><type class="union"><type>DateTimeZone</type><type>null</type></type><parameter>timezone</parameter><initializer>&null;</initializer></methodparam>
  </methodsynopsis>
  <para>&style.procedural;</para>
  <methodsynopsis>
   <type class="union"><type>DateTimeImmutable</type><type>false</type></type><methodname>date_create_immutable_from_format</methodname>
   <methodparam><type>string</type><parameter>format</parameter></methodparam>
   <methodparam><type>string</type><parameter>datetime</parameter></methodparam>
   <methodparam choice="opt"><type class="union"><type>DateTimeZone</type><type>null</type></type><parameter>timezone</parameter><initializer>&null;</initializer></methodparam>
  </methodsynopsis>
  <para>
   Returns a new DateTimeImmutable object representing the date and time specified by the
   <parameter>datetime</parameter> string, which was formatted in the given
   <parameter>format</parameter>.
  </para>
 </refsect1>

 <refsect1 role="parameters" xml:id="datetimeimmutable.createfromformat.parameters">
  &reftitle.parameters;
  <variablelist>
   <varlistentry>
    <term><parameter>format</parameter></term>
    <listitem>
     <para>
      The format that the passed in <type>string</type> should be in. See the
      formatting options below. In most cases, the same letters as for the
      <function>date</function> can be used.
     </para>
     <para>
      All fields are initialised with the current date/time. In most cases you
      would want to reset these to "zero" (the Unix epoch, <literal>1970-01-01
      00:00:00 UTC</literal>). You do that by including the
      <literal>!</literal> character as first character in your
      <parameter>format</parameter>, or <literal>|</literal> as your last.
      Please see the documentation for each character below for more
      information.
     </para>
     <para>
      The format is parsed from left to right, which means that in some
      situations the order in which the format characters are present affects
      the result. In the case of <literal>z</literal> (the day of the year),
      it is required that a year has already been parsed,
      for example through the <literal>Y</literal> or <literal>y</literal>
      characters.
     </para>
     <para>
      Letters that are used for parsing numbers allow a wide range of values,
      outside of what the logical range would be. For example, the
      <literal>d</literal> (day of the month) accepts values in the range from
      <literal>00</literal> to <literal>99</literal>. The only constraint is
      on the amount of digits. The date/time parser's overflow mechanism is
      used when out-of-range values are given. The examples below show some of
      this behaviour.
     </para>
     <para>
      This also means that the data parsed for a format letter is greedy, and
      will read up to the amount of digits its format allows for. That can
      then also mean that there are no
      longer enough characters in the <parameter>datetime</parameter> string
      for following format characters. An example on this page also
      illustrates this issue.
     </para>
     <para>
      <table>
       <title>The following characters are recognized in the
       <parameter>format</parameter> parameter string</title>
       <tgroup cols="3">
        <thead>
         <row>
          <entry><parameter>format</parameter> character</entry>
          <entry>Description</entry>
          <entry>Example parsable values</entry>
         </row>
        </thead>
        <tbody>
         <row>
          <entry align="center"><emphasis>Day</emphasis></entry>
          <entry>---</entry>
          <entry>---</entry>
         </row>
         <row>
          <entry><literal>d</literal> and <literal>j</literal></entry>
          <entry>Day of the month, 2 digits with or without leading zeros</entry>
          <entry>
           <literal>01</literal> to <literal>31</literal> or
           <literal>1</literal> to <literal>31</literal>. (2 digit numbers
           higher than the number of days in the month are accepted, in which
           case they will make the month overflow. For example using 33 with
           January, means February 2nd)
          </entry>
         </row>
         <row>
          <entry><literal>D</literal> and <literal>l</literal></entry>
          <entry>A textual representation of a day</entry>
          <entry>
           <literal>Mon</literal> through <literal>Sun</literal> or
           <literal>Sunday</literal> through <literal>Saturday</literal>. If
           the day name given is different then the day name belonging to a
           parsed (or default) date is different, then an overflow occurs to
           the <emphasis>next</emphasis> date with the given day name. See the
           examples below for an explanation.
          </entry>
         </row>
         <row>
          <entry><literal>S</literal></entry>
          <entry>English ordinal suffix for the day of the month, 2
          characters. It's ignored while processing.</entry>
          <entry>
           <literal>st</literal>, <literal>nd</literal>, <literal>rd</literal> or
           <literal>th</literal>.
          </entry>
         </row>
         <row>
          <entry><literal>z</literal></entry>
          <entry>
           The day of the year (starting from 0);
           must be preceded by <literal>Y</literal> or <literal>y</literal>.
          </entry>
          <entry>
           <literal>0</literal> through <literal>365</literal>. (3 digit
           numbers higher than the numbers in a year are accepted, in which
           case they will make the year overflow. For example using 366 with
           2022, means January 2nd, 2023)
          </entry>
         </row>
         <row>
          <entry align="center"><emphasis>Month</emphasis></entry>
          <entry>---</entry>
          <entry>---</entry>
         </row>
         <row>
          <entry><literal>F</literal> and <literal>M</literal></entry>
          <entry>A textual representation of a month, such as January or Sept</entry>
          <entry>
           <literal>January</literal> through <literal>December</literal> or
           <literal>Jan</literal> through <literal>Dec</literal>
          </entry>
         </row>
         <row>
          <entry><literal>m</literal> and <literal>n</literal></entry>
          <entry>Numeric representation of a month, with or without leading zeros</entry>
          <entry>
           <literal>01</literal> through <literal>12</literal> or
           <literal>1</literal> through <literal>12</literal>.
           (2 digit numbers higher than 12 are accepted, in which case they
           will make the year overflow. For example using 13 means January in
           the next year)
          </entry>
         </row>
         <row>
          <entry align="center"><emphasis>Year</emphasis></entry>
          <entry>---</entry>
          <entry>---</entry>
         </row>
         <row>
          <entry><literal>X</literal> and <literal>x</literal></entry>
          <entry>A full numeric representation of a year, up to 19 digits,
           optionally prefixed by <literal>+</literal> or
           <literal>-</literal></entry>
          <entry>Examples: <literal>0055</literal>, <literal>787</literal>,
           <literal>1999</literal>, <literal>-2003</literal>,
           <literal>+10191</literal></entry>
         </row>
         <row>
          <entry><literal>Y</literal></entry>
          <entry>A full numeric representation of a year, up to 4 digits</entry>
          <entry>Examples: <literal>0055</literal>, <literal>787</literal>,
           <literal>1999</literal>, <literal>2003</literal></entry>
         </row>
         <row>
          <entry><literal>y</literal></entry>
          <entry>
           A two digit representation of a year (which is assumed to be in the
           range 1970-2069, inclusive)
          </entry>
          <entry>
           Examples:
           <literal>99</literal> or <literal>03</literal>
           (which will be interpreted as <literal>1999</literal> and
           <literal>2003</literal>, respectively)
          </entry>
         </row>
         <row>
          <entry align="center"><emphasis>Time</emphasis></entry>
          <entry>---</entry>
          <entry>---</entry>
         </row>
         <row>
          <entry><literal>a</literal> and <literal>A</literal></entry>
          <entry>Ante meridiem and Post meridiem</entry>
          <entry><literal>am</literal> or <literal>pm</literal></entry>
         </row>
         <row>
          <entry><literal>g</literal> and <literal>h</literal></entry>
          <entry>12-hour format of an hour with or without leading zero</entry>
          <entry>
           <literal>1</literal> through <literal>12</literal> or
           <literal>01</literal> through <literal>12</literal> (2 digit
           numbers higher than 12 are accepted, in which case they will make
           the day overflow. For example using <literal>14</literal> means
           <literal>02</literal> in the next AM/PM period)
          </entry>
         </row>
         <row>
          <entry><literal>G</literal> and <literal>H</literal></entry>
          <entry>24-hour format of an hour with or without leading zeros</entry>
          <entry>
           <literal>0</literal> through <literal>23</literal> or
           <literal>00</literal> through <literal>23</literal> (2 digit
           numbers higher than 24 are accepted, in which case they will make
           the day overflow. For example using <literal>26</literal> means
           <literal>02:00</literal> the next day)
          </entry>
         </row>
         <row>
          <entry><literal>i</literal></entry>
          <entry>Minutes with leading zeros</entry>
          <entry>
           <literal>00</literal> to <literal>59</literal>. (2 digit
           numbers higher than 59 are accepted, in which case they will make
           the hour overflow. For example using <literal>66</literal> means
           <literal>:06</literal> the next hour)
          </entry>
         </row>
         <row>
          <entry><literal>s</literal></entry>
          <entry>Seconds, with leading zeros</entry>
          <entry>
           <literal>00</literal> through <literal>59</literal> (2 digit
           numbers higher than 59 are accepted, in which case they will make
           the minute overflow. For example using <literal>90</literal> means
           <literal>:30</literal> the next minute)
          </entry>
         </row>
         <row>
          <entry><literal>v</literal></entry>
          <entry>Fraction in milliseconds (up to three digits)</entry>
          <entry>Example: <literal>12</literal> (<literal>0.12</literal>
          seconds), <literal>345</literal> (<literal>0.345</literal> seconds)</entry>
         </row>
         <row>
          <entry><literal>u</literal></entry>
          <entry>Fraction in microseconds (up to six digits)</entry>
          <entry>Example: <literal>45</literal> (<literal>0.45</literal>
          seconds), <literal>654321</literal> (<literal>0.654321</literal>
          seconds)</entry>
         </row>
         <row>
          <entry align="center"><emphasis>Timezone</emphasis></entry>
          <entry>---</entry>
          <entry>---</entry>
         </row>
         <row>
          <entry>
           <literal>e</literal>, <literal>O</literal>, <literal>p</literal>,
           <literal>P</literal> and <literal>T</literal>
          </entry>
          <entry>Timezone identifier, or difference to UTC in hours, or
          difference to UTC with colon between hours and minutes, or timezone
          abbreviation</entry>
          <entry>Examples: <literal>UTC</literal>, <literal>GMT</literal>,
           <literal>Atlantic/Azores</literal> or
           <literal>+0200</literal> or <literal>+02:00</literal> or
           <literal>EST</literal>, <literal>MDT</literal>
          </entry>
         </row>
         <row>
          <entry align="center"><emphasis>Full Date/Time</emphasis></entry>
          <entry>---</entry>
          <entry>---</entry>
         </row>
         <row>
          <entry><literal>U</literal></entry>
          <entry>Seconds since the Unix Epoch (January 1 1970 00:00:00 GMT)</entry>
          <entry>Example: <literal>1292177455</literal></entry>
         </row>
         <row>
          <entry align="center"><emphasis>Whitespace and Separators</emphasis></entry>
          <entry>---</entry>
          <entry>---</entry>
         </row>
         <row>
          <entry><literal> </literal> (space)</entry>
          <entry>Zero or more spaces, tabs, NBSP (U+A0), or NNBSP (U+202F)
          characters</entry>
          <entry>Example: <literal>"\t"</literal>, <literal>"  "</literal></entry>
         </row>
         <row>
          <entry><literal>#</literal></entry>
          <entry>
           One of the following separation symbol: <literal>;</literal>,
           <literal>:</literal>, <literal>/</literal>, <literal>.</literal>,
           <literal>,</literal>, <literal>-</literal>, <literal>(</literal> or
           <literal>)</literal>
          </entry>
          <entry>Example: <literal>/</literal></entry>
         </row>
         <row>
          <entry>
           <literal>;</literal>,
           <literal>:</literal>, <literal>/</literal>, <literal>.</literal>,
           <literal>,</literal>, <literal>-</literal>, <literal>(</literal> or
           <literal>)</literal>
          </entry>
          <entry>The specified character.</entry>
          <entry>Example: <literal>-</literal></entry>
         </row>
         <row>
          <entry><literal>?</literal></entry>
          <entry>A random byte</entry>
          <entry>Example: <literal>^</literal> (Be aware that for UTF-8
          characters you might need more than one <literal>?</literal>.
          In this case, using <literal>*</literal> is probably what you want
          instead)</entry>
         </row>
         <row>
          <entry><literal>*</literal></entry>
          <entry>Random bytes until the next separator or digit</entry>
          <entry>Example: <literal>*</literal> in <literal>Y-*-d</literal> with
          the string <literal>2009-aWord-08</literal> will match
          <literal>aWord</literal></entry>
         </row>
         <row>
          <entry><literal>!</literal></entry>
          <entry>Resets all fields (year, month, day, hour, minute, second,
          fraction and timezone information) to zero-like values (
           <literal>0</literal> for hour, minute, second and fraction,
           <literal>1</literal> for month and day, <literal>1970</literal>
           for year and <literal>UTC</literal> for timezone information)</entry>
          <entry>Without <literal>!,</literal> all fields will be set to the
          current date and time.</entry>
         </row>
         <row>
          <entry><literal>|</literal></entry>
          <entry>Resets all fields (year, month, day, hour, minute, second,
          fraction and timezone information) to zero-like values if they have
          not been parsed yet</entry>
          <entry><literal>Y-m-d|</literal> will set the year, month and day
          to the information found in the string to parse, and sets the hour,
          minute and second to 0.</entry>
         </row>
         <row>
          <entry><literal>+</literal></entry>
          <entry>If this format specifier is present, trailing data in the
          string will not cause an error, but a warning instead</entry>
          <entry>Use <methodname>DateTimeImmutable::getLastErrors</methodname> to find out
          whether trailing data was present.</entry>
         </row>
        </tbody>
       </tgroup>
      </table>
     </para>
     <para>
      Unrecognized characters in the format string will cause the
      parsing to fail and an error message is appended to the returned
      structure. You can query error messages with
      <methodname>DateTimeImmutable::getLastErrors</methodname>.
     </para>
     <para>
      To include literal characters in <parameter>format</parameter>, you have
      to escape them with a backslash (<literal>\</literal>).
     </para>
     <para>
      If <parameter>format</parameter> does not contain the character
      <literal>!</literal> then portions of the generated date/time which are not
      specified in <parameter>format</parameter> will be set to the current
      system time.
     </para>
     <para>
      If <parameter>format</parameter> contains the
      character <literal>!</literal>, then portions of the generated
      date/time not provided in <parameter>format</parameter>, as well as
      values to the left-hand side of the <literal>!</literal>, will
      be set to corresponding values from the Unix epoch.
     </para>
     <para>
      If any time character is parsed, then all other time-related fields are
      set to "0", unless also parsed.
     </para>
     <para>
      The Unix epoch is 1970-01-01 00:00:00 UTC.
     </para>
    </listitem>
   </varlistentry>
   <varlistentry>
    <term><parameter>datetime</parameter></term>
    <listitem>
     <para>
      String representing the time.
     </para>
    </listitem>
   </varlistentry>
   <varlistentry>
    <term><parameter>timezone</parameter></term>
    <listitem>
     <para>
      A <classname>DateTimeZone</classname> object representing the
      desired time zone.
     </para>
     <para>
      If <parameter>timezone</parameter> is omitted or &null; and
      <parameter>datetime</parameter> contains no timezone,
      the current timezone will be used.
     </para>
     <note>
      <para>
       The <parameter>timezone</parameter> parameter
       and the current timezone are ignored when the
       <parameter>datetime</parameter> parameter either
       contains a UNIX timestamp (e.g. <literal>946684800</literal>)
       or specifies a timezone
       (e.g. <literal>2010-01-28T15:00:00+02:00</literal>).
      </para>
     </note>
    </listitem>
   </varlistentry>
  </variablelist>
 </refsect1>

 <refsect1 role="returnvalues">
  &reftitle.returnvalues;
  <para>
   Returns a new DateTimeImmutable instance&return.falseforfailure;.
  </para>
 </refsect1>

 <refsect1 role="errors">
  &reftitle.errors;
  <para>
   This method throws <exceptionname>ValueError</exceptionname> when the
   <parameter>datetime</parameter> contains NULL-bytes.
  </para>
 </refsect1>

 <refsect1 role="changelog">
  &reftitle.changelog;
  <para>
   <informaltable>
    <tgroup cols="2">
     <thead>
      <row>
       <entry>&Version;</entry>
       <entry>&Description;</entry>
      </row>
     </thead>
     <tbody>
      <row>
       <entry>8.2.9</entry>
       <entry>
        The <literal> </literal> (space) specifier now also supports NBSP
        (U+A0) and NNBSP (U+202F) characters.
       </entry>
      </row>
      <row>
       <entry>8.2.0</entry>
       <entry>
        The <literal>X</literal> and <literal>x</literal>
        <parameter>format</parameter> specifiers have been added.
       </entry>
      </row>
      <row>
       <entry>8.0.21, 8.1.8, 8.2.0</entry>
       <entry>
        Now throws <exceptionname>ValueError</exceptionname> when NULL-bytes
        are passed into <parameter>datetime</parameter>, which previously was silently
        ignored.
       </entry>
      </row>
      <row>
       <entry>7.3.0</entry>
       <entry>
        The <literal>v</literal> <parameter>format</parameter> specifier has
        been added.
       </entry>
      </row>
     </tbody>
    </tgroup>
   </informaltable>
  </para>
 </refsect1>

 <refsect1 role="examples">
  &reftitle.examples;
  <example>
   <title><function>DateTimeImmutable::createFromFormat</function> example</title>
   <para>&style.oop;</para>
   <programlisting role="php">
<![CDATA[
<?php
$date = DateTimeImmutable::createFromFormat('j-M-Y', '15-Feb-2009');
echo $date->format('Y-m-d');
?>
]]>
   </programlisting>
  </example>

  <example>
   <title>Using predefined format constants with <function>DateTimeImmutable::createFromFormat</function></title>
   <para>&style.oop;</para>
   <programlisting role="php">
<![CDATA[
<?php
$date = DateTimeImmutable::createFromFormat(DateTimeInterface::ISO8601, '2004-02-12T15:19:21+00:00');
$date = DateTimeImmutable::createFromFormat(DateTimeInterface::RFC3339_EXTENDED, '2013-10-14T09:00:00.000+02:00');
?>
]]>
   </programlisting>
   <para>
    The <link linkend="datetimeinterface.constants.types">formatting constants</link>
    as used in this example consist of
    a string of characters for <link linkend="datetime.format">formatting</link> a
    <classname>DateTimeImmutable</classname> object. In most cases, these
    letters match with the same elements of date/time information as the ones defined in the <link
    linkend="datetimeimmutable.createfromformat.parameters">parameters</link>
    section above, but they tend to be more lenient.
   </para>
  </example>

  <example>
   <title>Intricacies of <function>DateTimeImmutable::createFromFormat</function></title>
   <programlisting role="php">
<![CDATA[
<?php
echo 'Current time: ' . date('Y-m-d H:i:s') . "\n";

$format = 'Y-m-d';
$date = DateTimeImmutable::createFromFormat($format, '2009-02-15');
echo "Format: $format; " . $date->format('Y-m-d H:i:s') . "\n";

$format = 'Y-m-d H:i:s';
$date = DateTimeImmutable::createFromFormat($format, '2009-02-15 15:16:17');
echo "Format: $format; " . $date->format('Y-m-d H:i:s') . "\n";

$format = 'Y-m-!d H:i:s';
$date = DateTimeImmutable::createFromFormat($format, '2009-02-15 15:16:17');
echo "Format: $format; " . $date->format('Y-m-d H:i:s') . "\n";

$format = '!d';
$date = DateTimeImmutable::createFromFormat($format, '15');
echo "Format: $format; " . $date->format('Y-m-d H:i:s') . "\n";

$format = 'i';
$date = DateTimeImmutable::createFromFormat($format, '15');
echo "Format: $format; " . $date->format('Y-m-d H:i:s') . "\n";
?>
]]>
   </programlisting>
   &example.outputs.similar;
   <screen>
<![CDATA[
Current time: 2022-06-02 15:50:46
Format: Y-m-d; 2009-02-15 15:50:46
Format: Y-m-d H:i:s; 2009-02-15 15:16:17
Format: Y-m-!d H:i:s; 1970-01-15 15:16:17
Format: !d; 1970-01-15 00:00:00
Format: i; 2022-06-02 00:15:00
]]>
   </screen>
  </example>

  <example>
   <title>Format string with literal characters</title>
   <programlisting role="php">
<![CDATA[
<?php
echo DateTimeImmutable::createFromFormat('H\h i\m s\s','23h 15m 03s')->format('H:i:s');
?>
]]>
   </programlisting>
   &example.outputs.similar;
   <screen>
<![CDATA[
23:15:03
]]>
   </screen>
  </example>

  <example>
   <title>Overflow behaviour</title>
   <programlisting role="php">
<![CDATA[
<?php
echo DateTimeImmutable::createFromFormat('Y-m-d H:i:s', '2021-17-35 16:60:97')->format(DateTimeImmutable::RFC2822);
?>
]]>
   </programlisting>
   &example.outputs.similar;
   <screen>
<![CDATA[
Sat, 04 Jun 2022 17:01:37 +0000
]]>
   </screen>
   <para>
    Although the result looks odd, it is correct, as the following overflows
    happen:
   </para>
   <orderedlist>
    <listitem>
     <simpara>
      <literal>97</literal> seconds overflows to <literal>1</literal> minute,
      leaving <literal>37</literal> seconds.
     </simpara>
    </listitem>
    <listitem>
     <simpara>
      <literal>61</literal> minutes overflows to <literal>1</literal> hour,
      leaving <literal>1</literal> minutes.
     </simpara>
    </listitem>
    <listitem>
     <simpara>
      <literal>35</literal> days overflows to <literal>1</literal> month,
      leaving <literal>4</literal> days. The amount of days that are left over
      depends on the month, as not every month has the same amount of days.
     </simpara>
    </listitem>
    <listitem>
     <simpara>
      <literal>18</literal> months overflows to <literal>1</literal> year,
      leaving <literal>6</literal> months.
     </simpara>
    </listitem>
   </orderedlist>
  </example>

  <example>
   <title>Overflowing day name behaviour</title>
   <programlisting role="php">
<![CDATA[
<?php
$d = DateTime::createFromFormat(DateTimeInterface::RFC1123, 'Mon, 3 Aug 2020 25:00:00 +0000');
echo $d->format(DateTime::RFC1123), "\n";
?>
]]>
   </programlisting>
   &example.outputs.similar;
   <screen>
<![CDATA[
Mon, 10 Aug 2020 01:00:00 +0000
]]>
   </screen>
   <para>
    Although the result looks odd, it is correct, as the following overflows
    happen:
   </para>
   <orderedlist>
    <listitem>
     <simpara>
      <literal>3 Aug 2020 25:00:00</literal> overflows to <literal>(Tue) 4 Aug
      2020 01:00</literal>.
     </simpara>
    </listitem>
    <listitem>
     <simpara>
      <literal>Mon</literal> gets applied, which advances the date to
      <literal>Mon, 10 Aug 2020 01:00:00</literal>. The explanation of
      relative keywords such as <literal>Mon</literal> is explained in the
      section on <link linkend="datetime.formats.relative">relative
      formats</link>.
     </simpara>
    </listitem>
   </orderedlist>
  </example>

  <para>
   In order to detect overflows in dates, you can use
   <methodname>DateTimeImmutable::getLastErrors</methodname>, which will
   include a warning if an overflow occured.
  </para>
  <example>
   <title>Detecting overflown dates</title>
   <programlisting role="php">
<![CDATA[
<?php
$d = DateTimeImmutable::createFromFormat('Y-m-d H:i:s', '2021-17-35 16:60:97');
echo $d->format(DateTimeImmutable::RFC2822), "\n\n";

var_dump(DateTimeImmutable::GetLastErrors());
?>
]]>
   </programlisting>
   &example.outputs.similar;
   <screen>
<![CDATA[
Sat, 04 Jun 2022 17:01:37 +0000

array(4) {
  'warning_count' =>
  int(2)
  'warnings' =>
  array(1) {
    [19] =>
    string(27) "The parsed date was invalid"
  }
  'error_count' =>
  int(0)
  'errors' =>
  array(0) {
  }
}
]]>
   </screen>
  </example>

  <example>
   <title>Greedy parsing behaviour</title>
   <programlisting role="php">
<![CDATA[
<?php
print_r(date_parse_from_format('Gis', '60101'));
?>
]]>
   </programlisting>
   &example.outputs.similar;
   <screen>
<![CDATA[
Array
(
    [year] =>
    [month] =>
    [day] =>
    [hour] => 60
    [minute] => 10
    [second] => 0
    [fraction] => 0
    [warning_count] => 1
    [warnings] => Array
        (
            [5] => The parsed time was invalid
        )

    [error_count] => 1
    [errors] => Array
        (
            [4] => A two digit second could not be found
        )

    [is_localtime] =>
)
]]>
   </screen>
   <para>
    The <literal>G</literal> format is to parse 24 hour clock hours, with or
    without leading zero. This requires to parse 1 or 2 digits. Because there
    are two following digits, it greedily reads this as <literal>60</literal>.
   </para>
   <para>
    The following <literal>i</literal> and <literal>s</literal> format
    characters both require two digits. This means that <literal>10</literal>
    is passed as minute (<literal>i</literal>), and that there are then not
    enough digits left to parse for as second (<literal>s</literal>).
   </para>
   <para>
    The <literal>errors</literal> array indicates this problem.
   </para>
   <para>
    Additionally, an hour of <literal>60</literal> is outside the range
    <literal>0</literal>-<literal>24</literal>, which causes the
    <literal>warnings</literal> array to include a warning that the time is
    invalid.
   </para>
  </example>

 </refsect1>

 <refsect1 role="seealso">
  &reftitle.seealso;
  <simplelist>
   <member><function>DateTimeImmutable::__construct</function></member>
   <member><function>DateTimeImmutable::getLastErrors</function></member>
   <member><function>checkdate</function></member>
   <member><function>strptime</function></member>
  </simplelist>
 </refsect1>

</refentry>
<!-- Keep this comment at the end of the file
Local variables:
mode: sgml
sgml-omittag:t
sgml-shorttag:t
sgml-minimize-attributes:nil
sgml-always-quote-attributes:t
sgml-indent-step:1
sgml-indent-data:t
indent-tabs-mode:nil
sgml-parent-document:nil
sgml-default-dtd-file:"~/.phpdoc/manual.ced"
sgml-exposed-tags:nil
sgml-local-catalogs:nil
sgml-local-ecat-files:nil
End:
vim600: syn=xml fen fdm=syntax fdl=2 si
vim: et tw=78 syn=sgml
vi: ts=1 sw=1
-->