_chapter 4.htm

linux、unix初学者的必读书籍 详细讲述了shell编程方法与技巧

    </td>
    <td class="docTableCell" vAlign="top">Globally substitutes on a line. </td>
  </tr>
  <tr>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">p</span>
    </td>
    <td class="docTableCell" vAlign="top">Prints lines. </td>
  </tr>
  <tr>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">w</span>
    </td>
    <td class="docTableCell" vAlign="top">Writes lines out to a file. </td>
  </tr>
  <tr>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">x</span>
    </td>
    <td class="docTableCell" vAlign="top">Exchanges contents of the holding
    buffer with the pattern space. </td>
  </tr>
  <tr>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">y</span>
    </td>
    <td class="docTableCell" vAlign="top">Translates one character to another
    (cannot use regular expression metacharacters with <span class="docEmphasis">
    y</span>). </td>
  </tr>
</table>
<p>&nbsp;</p>
<table cellSpacing="0" cellPadding="1" width="100%" border="1">
  <caption>
  <h5 id="ch04table02" class="docTableTitle">Table 4.2. <span class="docEmphasis">sed</span>
  Options</h5>
  </caption>
  <colgroup span="2" align="left">
  </colgroup>
  <tr>
    <th class="docTableHeader" vAlign="top"><span class="docEmphBoldItalic">
    Options</span> </th>
    <th class="docTableHeader" vAlign="top"><span class="docEmphBoldItalic">
    Function</span> </th>
  </tr>
  <tr>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">-e</span>
    </td>
    <td class="docTableCell" vAlign="top">Allows multiple edits. </td>
  </tr>
  <tr>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">-f</span>
    </td>
    <td class="docTableCell" vAlign="top">Precedes a <span class="docEmphasis">
    sed</span> script filename. </td>
  </tr>
  <tr>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">-n</span>
    </td>
    <td class="docTableCell" vAlign="top">Suppresses default output. </td>
  </tr>
</table>
<p class="docText">When multiple commands are used or addresses need to be
nested within a range of addresses, the commands are enclosed in curly braces
and each command is either on a separate line or terminated with semicolons.</p>
<p class="docText">The exclamation point (!) can be used to negate a command.
For example,</p>
<pre>% <span class="docEmphStrong">sed '/Tom/d' file</span></pre>
<p class="docText">tells <span class="docEmphasis">sed</span> to delete all
lines containing the pattern <span class="docEmphasis">Tom,</span> whereas</p>
<pre>% <span class="docEmphStrong">sed '/Tom/!d' file</span>
</pre>
<p class="docText">tells <span class="docEmphasis">sed</span> to delete lines
not containing <span class="docEmphasis">Tom.</span></p>
<p class="docText">The <span class="docEmphasis">sed</span> options are
<span class="docEmphasis">-e, -f,</span> and, <span class="docEmphasis">-n.</span>
The <span class="docEmphasis">-e</span> is used for multiple edits at the
command line, the <span class="docEmphasis">-f</span> precedes a
<span class="docEmphasis">sed</span> script filename, and the
<span class="docEmphasis">-n</span> suppresses printing output.</p>
<h3 class="docSection1Title" id="ch04lev1sec5">4.5 Error Messages and Exit Status</h3>
<p class="docText">When <span class="docEmphasis">sed</span> encounters a syntax
error, it sends a pretty straightforward error message to standard error, but if
it cannot figure out what you did wrong, <span class="docEmphasis">sed</span>
gets &quot;garbled,&quot; which we could guess means confused. The exit status that
<span class="docEmphasis">sed</span> returns to the shell, if its syntax is
error-free, is a zero for success and a nonzero integer for failure.<span id="ENB4-2"><a class="docLink" href="#EN4-2"><sup>[2]</sup></a></span></p>
<h5 id="ch04list03" class="docExampleTitle">Example 4.3 </h5>
<pre>1   % <span class="docEmphStrong">sed '1,3v ' file</span>
    s<span class="docEmphasis">ed: Unrecognized command: 1,3v</span>
    % <span class="docEmphStrong">echo $status</span> (<span class="docEmphStrong">echo $?</span>  <span class="docEmphasis">if using Korn or Bourne shell</span>)
    <span class="docEmphasis">2</span>

2   % <span class="docEmphStrong">sed '/^John' file</span>
    <span class="docEmphasis">sed: Illegal or missing delimiter: /^John</span>

3   % <span class="docEmphStrong">sed 's/134345/g' file</span>
    <span class="docEmphasis">sed: Ending delimiter missing on substitution: s/134345/g</span></pre>
<table cellSpacing="0" width="90%" border="1" align="center">
  <tr>
    <td>
    <h2 class="docSidebarTitle">EXPLANATION</h2>
    <span style="FONT-WEIGHT: bold">
    <ol class="docList" type="1">
      <li><span style="FONT-WEIGHT: normal">
      <p class="docList">The <span class="docEmphasis">v</span> command is
      unrecognized by <span class="docEmphasis">sed.</span> The exit status was
      2, indicating that <span class="docEmphasis">sed</span> exited with a
      syntax problem.</span></li>
      <li><span style="FONT-WEIGHT: normal">
      <p class="docList">The pattern <span class="docEmphasis">/^John</span> is
      missing the closing forward slash.</span></li>
      <li><span style="FONT-WEIGHT: normal">
      <p class="docList">The substitution command, <span class="docEmphasis">s,</span>
      contains the search string but not the replacement string.</span></li>
    </ol>
    </span></td>
  </tr>
</table>
<h4 class="docSection2Title" id="ch04lev2sec1">4.5.1 Metacharacters</h4>
<p class="docText">Like <span class="docEmphasis">grep, sed</span> supports a
number of special metacharacters to control pattern searching. See
<a class="docLink" href="#ch04table03">Table 4.3</a></p>
<table cellSpacing="0" cellPadding="1" width="100%" border="1">
  <caption>
  <h5 id="ch04table03" class="docTableTitle">Table 4.3. <span class="docEmphasis">sed</span>'s
  Regular Expression Metacharacters.</h5>
  </caption>
  <colgroup span="4" align="left">
  </colgroup>
  <tr>
    <th class="docTableHeader" vAlign="top"><span class="docEmphBoldItalic">
    Metacharacter</span> </th>
    <th class="docTableHeader" vAlign="top"><span class="docEmphBoldItalic">
    Function</span> </th>
    <th class="docTableHeader" vAlign="top"><span class="docEmphBoldItalic">
    Example</span> </th>
    <th class="docTableHeader" vAlign="top"><span class="docEmphBoldItalic">What
    It Matches</span> </th>
  </tr>
  <tr>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">^</span>
    </td>
    <td class="docTableCell" vAlign="top">Beginning-of-line anchor </td>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">/^love/</span>
    </td>
    <td class="docTableCell" vAlign="top">Matches all lines beginning with
    <span class="docEmphasis">love.</span> </td>
  </tr>
  <tr>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">$</span>
    </td>
    <td class="docTableCell" vAlign="top">End-of-line anchor </td>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">/love$/</span>
    </td>
    <td class="docTableCell" vAlign="top">Matches all lines ending with
    <span class="docEmphasis">love.</span> </td>
  </tr>
  <tr>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">.</span></td>
    <td class="docTableCell" vAlign="top">Matches one character, but not the
    newline character </td>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">/l..e/</span>
    </td>
    <td class="docTableCell" vAlign="top">Matches lines containing an
    <span class="docEmphasis">l,</span> followed by two characters, followed by
    an <span class="docEmphasis">e.</span> </td>
  </tr>
  <tr>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">*</span>
    </td>
    <td class="docTableCell" vAlign="top">Matches zero or more characters </td>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">/ *love/</span>
    </td>
    <td class="docTableCell" vAlign="top">Matches lines with zero or more
    spaces, followed by the pattern <span class="docEmphasis">love.</span> </td>
  </tr>
  <tr>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">[ ]</span>
    </td>
    <td class="docTableCell" vAlign="top">Matches one character in the set </td>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">/[Ll]ove/</span>
    </td>
    <td class="docTableCell" vAlign="top">Matches lines containing
    <span class="docEmphasis">love</span> or <span class="docEmphasis">Love.</span>
    </td>
  </tr>
  <tr>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">[^ ]</span>
    </td>
    <td class="docTableCell" vAlign="top">Matches one character not in the set
    </td>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">/[^A-KM-Z]ove/</span>
    </td>
    <td class="docTableCell" vAlign="top">Matches lines not containing
    <span class="docEmphasis">A</span> through <span class="docEmphasis">K</span>
    or <span class="docEmphasis">M</span> through <span class="docEmphasis">Z</span>
    followed by <span class="docEmphasis">ove.</span> </td>
  </tr>
  <tr>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">\(..\)</span>
    </td>
    <td class="docTableCell" vAlign="top">Saves matched characters </td>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">
    s/\(love\)able/\1er/</span> </td>
    <td class="docTableCell" vAlign="top">Tags marked portion and saves it as
    tag number 1. To reference later, use <span class="docEmphasis">\1</span> to
    reference the pattern. May use up to nine tags, starting with the first tag
    at the leftmost part of the pattern. For example, <span class="docEmphasis">
    love</span> is saved in register 1 and remembered in the replacement string.
    <span class="docEmphasis">lovable</span> is replaced with
    <span class="docEmphasis">lover.</span> </td>
  </tr>
  <tr>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">&amp;</span>
    </td>
    <td class="docTableCell" vAlign="top">Saves search string so it can be
    remembered in the replacement string </td>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">
    s/love/**&amp;**/</span> </td>
    <td class="docTableCell" vAlign="top">The ampersand represents the search
    string. The string <span class="docEmphasis">love</span> will be replaced
    with itself surrounded by asterisks; i.e., <span class="docEmphasis">love</span>
    will become <span class="docEmphasis">**love**.</span> </td>
  </tr>
  <tr>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">\&lt;</span>
    </td>
    <td class="docTableCell" vAlign="top">Beginning-of-word anchor </td>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">/\&lt;love/</span>
    </td>
    <td class="docTableCell" vAlign="top">Matches lines containing a word that
    begins with <span class="docEmphasis">love.</span> </td>
  </tr>
  <tr>
    <td class="docTableCell" vAlign="top"><span class="docEmphasis">\&gt;</span>