0x1949 Team - FAZEMRX - MANAGER

Edit File: profile.html

<!DOCTYPE html>

<title>26.4. The Python Profilers &#8212; Python 2.7.18 documentation</title>
 <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
 <link rel="stylesheet" type="text/css" href="../_static/classic.css" />
 
 <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
 <script src="../_static/jquery.js"></script>
 <script src="../_static/underscore.js"></script>
 <script src="../_static/doctools.js"></script>
 
 <script src="../_static/sidebar.js"></script>
 
 <link rel="search" type="application/opensearchdescription+xml"
 title="Search within Python 2.7.18 documentation"
 href="../_static/opensearch.xml"/>
 <link rel="author" title="About these documents" href="../about.html" />
 <link rel="index" title="Index" href="../genindex.html" />
 <link rel="search" title="Search" href="../search.html" />
 <link rel="copyright" title="Copyright" href="../copyright.html" />
 <link rel="next" title="26.5. hotshot — High performance logging profiler" href="hotshot.html" />
 <link rel="prev" title="26.2. pdb — The Python Debugger" href="pdb.html" />
 <link rel="shortcut icon" type="image/png" href="../_static/py.png" />
 <link rel="canonical" href="file:///usr/share/doc/python2.7/html/library/profile.html" />
 <script type="text/javascript" src="../_static/copybutton.js"></script>

</head><body>
<div id="outdated-warning" style="padding: .5em; text-align: center; background-color: #FFBABA; color: #6A0E0E;">
 This document is for an old version of Python that is <a href="https://devguide.python.org/devcycle/#end-of-life-branches">no longer supported</a>.
 You should install the python3 and python3-doc packages and read the 
 <a href="file:///usr/share/doc/python3-doc/html/library/profile.html"> Python documentation for the Python3 version packaged in this release</a>.
</div>
 
 <div class="related" role="navigation" aria-label="related navigation">
 <h3>Navigation</h3>
 <ul>
 <li class="right" style="margin-right: 10px">
 <a href="../genindex.html" title="General Index"
 accesskey="I">index</a></li>
 <li class="right" >
 <a href="../py-modindex.html" title="Python Module Index"
 >modules</a> |</li>
 <li class="right" >
 <a href="hotshot.html" title="26.5. hotshot — High performance logging profiler"
 accesskey="N">next</a> |</li>
 <li class="right" >
 <a href="pdb.html" title="26.2. pdb — The Python Debugger"
 accesskey="P">previous</a> |</li>
 <li><img src="../_static/py.png" alt=""
 style="vertical-align: middle; margin-top: -1px"/></li>
 <li><a href="https://www.python.org/">Python</a> &#187;</li>
 <li>
 <a href="../index.html">Python 2.7.18 documentation</a> &#187;
 </li>

<li class="nav-item nav-item-1"><a href="index.html" >The Python Standard Library</a> &#187;</li>
 <li class="nav-item nav-item-2"><a href="debug.html" accesskey="U">26. Debugging and Profiling</a> &#187;</li>
 <li class="nav-item nav-item-this"><a href="">26.4. The Python Profilers</a></li> 
 </ul>
 </div>

<div class="document">
 <div class="documentwrapper">
 <div class="bodywrapper">
 <div class="body" role="main">
 
 <section id="the-python-profilers">
<h1>26.4. The Python Profilers<a class="headerlink" href="#the-python-profilers" title="Permalink to this headline">¶</a></h1>
Source code: <a class="reference external" href="https://github.com/python/cpython/tree/2.7/Lib/profile.py">Lib/profile.py</a> and <a class="reference external" href="https://github.com/python/cpython/tree/2.7/Lib/pstats.py">Lib/pstats.py</a>
<hr class="docutils" />
<section id="introduction-to-the-profilers">
<h2>26.4.1. Introduction to the profilers<a class="headerlink" href="#introduction-to-the-profilers" title="Permalink to this headline">¶</a></h2>
<a class="reference internal" href="#module-cProfile" title="cProfile"><code class="xref py py-mod docutils literal notranslate">cProfile</code></a> and <a class="reference internal" href="#module-profile" title="profile: Python source profiler."><code class="xref py py-mod docutils literal notranslate">profile</code></a> provide deterministic profiling of
Python programs. A profile is a set of statistics that describes how
often and for how long various parts of the program executed. These statistics
can be formatted into reports via the <a class="reference internal" href="#module-pstats" title="pstats: Statistics object for use with the profiler."><code class="xref py py-mod docutils literal notranslate">pstats</code></a> module.
The Python standard library provides three different implementations of the same
profiling interface:
<ol class="arabic">
<li><a class="reference internal" href="#module-cProfile" title="cProfile"><code class="xref py py-mod docutils literal notranslate">cProfile</code></a> is recommended for most users; it’s a C extension with
reasonable overhead that makes it suitable for profiling long-running
programs. Based on <code class="xref py py-mod docutils literal notranslate">lsprof</code>, contributed by Brett Rosen and Ted
Czotter.
<div class="versionadded">
New in version 2.5.
</div>
</li>
<li><a class="reference internal" href="#module-profile" title="profile: Python source profiler."><code class="xref py py-mod docutils literal notranslate">profile</code></a>, a pure Python module whose interface is imitated by
<a class="reference internal" href="#module-cProfile" title="cProfile"><code class="xref py py-mod docutils literal notranslate">cProfile</code></a>, but which adds significant overhead to profiled programs.
If you’re trying to extend the profiler in some way, the task might be easier
with this module. Originally designed and written by Jim Roskind.
<div class="versionchanged">
Changed in version 2.4: Now also reports the time spent in calls to built-in functions
and methods.
</div>
</li>
<li><a class="reference internal" href="hotshot.html#module-hotshot" title="hotshot: High performance logging profiler, mostly written in C."><code class="xref py py-mod docutils literal notranslate">hotshot</code></a> was an experimental C module that focused on minimizing
the overhead of profiling, at the expense of longer data
post-processing times. It is no longer maintained and may be
dropped in a future version of Python.
<div class="versionchanged">
Changed in version 2.5: The results should be more meaningful than in the past: the timing core
contained a critical bug.
</div>
</li>
</ol>
The <a class="reference internal" href="#module-profile" title="profile: Python source profiler."><code class="xref py py-mod docutils literal notranslate">profile</code></a> and <a class="reference internal" href="#module-cProfile" title="cProfile"><code class="xref py py-mod docutils literal notranslate">cProfile</code></a> modules export the same interface, so
they are mostly interchangeable; <a class="reference internal" href="#module-cProfile" title="cProfile"><code class="xref py py-mod docutils literal notranslate">cProfile</code></a> has a much lower overhead but
is newer and might not be available on all systems.
<a class="reference internal" href="#module-cProfile" title="cProfile"><code class="xref py py-mod docutils literal notranslate">cProfile</code></a> is really a compatibility layer on top of the internal
<code class="xref py py-mod docutils literal notranslate">_lsprof</code> module. The <a class="reference internal" href="hotshot.html#module-hotshot" title="hotshot: High performance logging profiler, mostly written in C."><code class="xref py py-mod docutils literal notranslate">hotshot</code></a> module is reserved for specialized
usage.
<div class="admonition note">
Note
The profiler modules are designed to provide an execution profile for a given
program, not for benchmarking purposes (for that, there is <a class="reference internal" href="timeit.html#module-timeit" title="timeit: Measure the execution time of small code snippets."><code class="xref py py-mod docutils literal notranslate">timeit</code></a> for
reasonably accurate results). This particularly applies to benchmarking
Python code against C code: the profilers introduce overhead for Python code,
but not for C-level functions, and so the C code would seem faster than any
Python one.
</div>
</section>
<section id="instant-user-s-manual">
<h2>26.4.2. Instant User’s Manual<a class="headerlink" href="#instant-user-s-manual" title="Permalink to this headline">¶</a></h2>
This section is provided for users that “don’t want to read the manual.” It
provides a very brief overview, and allows a user to rapidly perform profiling
on an existing application.
To profile a function that takes a single argument, you can do:
<div class="highlight-default notranslate"><div class="highlight"><pre>import cProfile
import re
cProfile.run(&#39;re.compile(&quot;foo|bar&quot;)&#39;)
</pre></div>
</div>
(Use <a class="reference internal" href="#module-profile" title="profile: Python source profiler."><code class="xref py py-mod docutils literal notranslate">profile</code></a> instead of <a class="reference internal" href="#module-cProfile" title="cProfile"><code class="xref py py-mod docutils literal notranslate">cProfile</code></a> if the latter is not available on
your system.)
The above action would run <a class="reference internal" href="re.html#re.compile" title="re.compile"><code class="xref py py-func docutils literal notranslate">re.compile()</code></a> and print profile results like
the following:
<div class="highlight-default notranslate"><div class="highlight"><pre> 197 function calls (192 primitive calls) in 0.002 seconds

Ordered by: standard name

ncalls tottime percall cumtime percall filename:lineno(function)
 1 0.000 0.000 0.001 0.001 &lt;string&gt;:1(&lt;module&gt;)
 1 0.000 0.000 0.001 0.001 re.py:212(compile)
 1 0.000 0.000 0.001 0.001 re.py:268(_compile)
 1 0.000 0.000 0.000 0.000 sre_compile.py:172(_compile_charset)
 1 0.000 0.000 0.000 0.000 sre_compile.py:201(_optimize_charset)
 4 0.000 0.000 0.000 0.000 sre_compile.py:25(_identityfunction)
 3/1 0.000 0.000 0.000 0.000 sre_compile.py:33(_compile)
</pre></div>
</div>
The first line indicates that 197 calls were monitored. Of those calls, 192
were primitive, meaning that the call was not induced via recursion. The
next line: <code class="docutils literal notranslate">Ordered by: standard name</code>, indicates that the text string in the
far right column was used to sort the output. The column headings include:
<dl class="simple">
<dt>ncalls</dt><dd>for the number of calls,
</dd>
<dt>tottime</dt><dd>for the total time spent in the given function (and excluding time made in
calls to sub-functions)
</dd>
<dt>percall</dt><dd>is the quotient of <code class="docutils literal notranslate">tottime</code> divided by <code class="docutils literal notranslate">ncalls</code>
</dd>
<dt>cumtime</dt><dd>is the cumulative time spent in this and all subfunctions (from invocation
till exit). This figure is accurate even for recursive functions.
</dd>
<dt>percall</dt><dd>is the quotient of <code class="docutils literal notranslate">cumtime</code> divided by primitive calls
</dd>
<dt>filename:lineno(function)</dt><dd>provides the respective data of each function
</dd>
</dl>
When there are two numbers in the first column (for example <code class="docutils literal notranslate">3/1</code>), it means
that the function recursed. The second value is the number of primitive calls
and the former is the total number of calls. Note that when the function does
not recurse, these two values are the same, and only the single figure is
printed.
Instead of printing the output at the end of the profile run, you can save the
results to a file by specifying a filename to the <code class="xref py py-func docutils literal notranslate">run()</code> function:
<div class="highlight-default notranslate"><div class="highlight"><pre>import cProfile
import re
cProfile.run(&#39;re.compile(&quot;foo|bar&quot;)&#39;, &#39;restats&#39;)
</pre></div>
</div>
The <a class="reference internal" href="#pstats.Stats" title="pstats.Stats"><code class="xref py py-class docutils literal notranslate">pstats.Stats</code></a> class reads profile results from a file and formats
them in various ways.
The file <a class="reference internal" href="#module-cProfile" title="cProfile"><code class="xref py py-mod docutils literal notranslate">cProfile</code></a> can also be invoked as a script to profile another
script. For example:
<div class="highlight-default notranslate"><div class="highlight"><pre>python -m cProfile [-o output_file] [-s sort_order] myscript.py
</pre></div>
</div>
<code class="docutils literal notranslate">-o</code> writes the profile results to a file instead of to stdout
<code class="docutils literal notranslate">-s</code> specifies one of the <a class="reference internal" href="#pstats.Stats.sort_stats" title="pstats.Stats.sort_stats"><code class="xref py py-func docutils literal notranslate">sort_stats()</code></a> sort values to sort
the output by. This only applies when <code class="docutils literal notranslate">-o</code> is not supplied.
The <a class="reference internal" href="#module-pstats" title="pstats: Statistics object for use with the profiler."><code class="xref py py-mod docutils literal notranslate">pstats</code></a> module’s <a class="reference internal" href="#pstats.Stats" title="pstats.Stats"><code class="xref py py-class docutils literal notranslate">Stats</code></a> class has a variety of methods
for manipulating and printing the data saved into a profile results file:
<div class="highlight-default notranslate"><div class="highlight"><pre>import pstats
p = pstats.Stats(&#39;restats&#39;)
p.strip_dirs().sort_stats(-1).print_stats()
</pre></div>
</div>
The <a class="reference internal" href="#pstats.Stats.strip_dirs" title="pstats.Stats.strip_dirs"><code class="xref py py-meth docutils literal notranslate">strip_dirs()</code></a> method removed the extraneous path from all
the module names. The <a class="reference internal" href="#pstats.Stats.sort_stats" title="pstats.Stats.sort_stats"><code class="xref py py-meth docutils literal notranslate">sort_stats()</code></a> method sorted all the
entries according to the standard module/line/name string that is printed. The
<a class="reference internal" href="#pstats.Stats.print_stats" title="pstats.Stats.print_stats"><code class="xref py py-meth docutils literal notranslate">print_stats()</code></a> method printed out all the statistics. You
might try the following sort calls:
<div class="highlight-default notranslate"><div class="highlight"><pre>p.sort_stats(&#39;name&#39;)
p.print_stats()
</pre></div>
</div>
The first call will actually sort the list by function name, and the second call
will print out the statistics. The following are some interesting calls to
experiment with:
<div class="highlight-default notranslate"><div class="highlight"><pre>p.sort_stats(&#39;cumulative&#39;).print_stats(10)
</pre></div>
</div>
This sorts the profile by cumulative time in a function, and then only prints
the ten most significant lines. If you want to understand what algorithms are
taking time, the above line is what you would use.
If you were looking to see what functions were looping a lot, and taking a lot
of time, you would do:
<div class="highlight-default notranslate"><div class="highlight"><pre>p.sort_stats(&#39;time&#39;).print_stats(10)
</pre></div>
</div>
to sort according to time spent within each function, and then print the
statistics for the top ten functions.
You might also try:
<div class="highlight-default notranslate"><div class="highlight"><pre>p.sort_stats(&#39;file&#39;).print_stats(&#39;__init__&#39;)
</pre></div>
</div>
This will sort all the statistics by file name, and then print out statistics
for only the class init methods (since they are spelled with <code class="docutils literal notranslate">__init__</code> in
them). As one final example, you could try:
<div class="highlight-default notranslate"><div class="highlight"><pre>p.sort_stats(&#39;time&#39;, &#39;cum&#39;).print_stats(.5, &#39;init&#39;)
</pre></div>
</div>
This line sorts statistics with a primary key of time, and a secondary key of
cumulative time, and then prints out some of the statistics. To be specific, the
list is first culled down to 50% (re: <code class="docutils literal notranslate">.5</code>) of its original size, then only
lines containing <code class="docutils literal notranslate">init</code> are maintained, and that sub-sub-list is printed.
If you wondered what functions called the above functions, you could now (<code class="docutils literal notranslate">p</code>
is still sorted according to the last criteria) do:
<div class="highlight-default notranslate"><div class="highlight"><pre>p.print_callers(.5, &#39;init&#39;)
</pre></div>
</div>
and you would get a list of callers for each of the listed functions.
If you want more functionality, you’re going to have to read the manual, or
guess what the following functions do:
<div class="highlight-default notranslate"><div class="highlight"><pre>p.print_callees()
p.add(&#39;restats&#39;)
</pre></div>
</div>
Invoked as a script, the <a class="reference internal" href="#module-pstats" title="pstats: Statistics object for use with the profiler."><code class="xref py py-mod docutils literal notranslate">pstats</code></a> module is a statistics browser for
reading and examining profile dumps. It has a simple line-oriented interface
(implemented using <a class="reference internal" href="cmd.html#module-cmd" title="cmd: Build line-oriented command interpreters."><code class="xref py py-mod docutils literal notranslate">cmd</code></a>) and interactive help.
</section>
<section id="module-cProfile">
<h2>26.4.3. <a class="reference internal" href="#module-profile" title="profile: Python source profiler."><code class="xref py py-mod docutils literal notranslate">profile</code></a> and <a class="reference internal" href="#module-cProfile" title="cProfile"><code class="xref py py-mod docutils literal notranslate">cProfile</code></a> Module Reference<a class="headerlink" href="#module-cProfile" title="Permalink to this headline">¶</a></h2>
Both the <a class="reference internal" href="#module-profile" title="profile: Python source profiler."><code class="xref py py-mod docutils literal notranslate">profile</code></a> and <a class="reference internal" href="#module-cProfile" title="cProfile"><code class="xref py py-mod docutils literal notranslate">cProfile</code></a> modules provide the following
functions:
<dl class="py function">
<dt class="sig sig-object py" id="profile.run">
profile.run(command, filename=None, sort=- 1)<a class="headerlink" href="#profile.run" title="Permalink to this definition">¶</a></dt>
<dd>This function takes a single argument that can be passed to the <code class="xref py py-func docutils literal notranslate">exec()</code>
function, and an optional file name. In all cases this routine executes:
<div class="highlight-default notranslate"><div class="highlight"><pre>exec(command, __main__.__dict__, __main__.__dict__)
</pre></div>
</div>
and gathers profiling statistics from the execution. If no file name is
present, then this function automatically creates a <a class="reference internal" href="#pstats.Stats" title="pstats.Stats"><code class="xref py py-class docutils literal notranslate">Stats</code></a>
instance and prints a simple profiling report. If the sort value is specified
it is passed to this <a class="reference internal" href="#pstats.Stats" title="pstats.Stats"><code class="xref py py-class docutils literal notranslate">Stats</code></a> instance to control how the
results are sorted.
</dd></dl>

<dl class="py function">
<dt class="sig sig-object py" id="profile.runctx">
profile.runctx(command, globals, locals, filename=None)<a class="headerlink" href="#profile.runctx" title="Permalink to this definition">¶</a></dt>
<dd>This function is similar to <a class="reference internal" href="#profile.run" title="profile.run"><code class="xref py py-func docutils literal notranslate">run()</code></a>, with added arguments to supply the
globals and locals dictionaries for the command string. This routine
executes:
<div class="highlight-default notranslate"><div class="highlight"><pre>exec(command, globals, locals)
</pre></div>
</div>
and gathers profiling statistics as in the <a class="reference internal" href="#profile.run" title="profile.run"><code class="xref py py-func docutils literal notranslate">run()</code></a> function above.
</dd></dl>

<dl class="py class">
<dt class="sig sig-object py" id="profile.Profile">
class profile.Profile(timer=None, timeunit=0.0, subcalls=True, builtins=True)<a class="headerlink" href="#profile.Profile" title="Permalink to this definition">¶</a></dt>
<dd>This class is normally only used if more precise control over profiling is
needed than what the <code class="xref py py-func docutils literal notranslate">cProfile.run()</code> function provides.
A custom timer can be supplied for measuring how long code takes to run via
the timer argument. This must be a function that returns a single number
representing the current time. If the number is an integer, the timeunit
specifies a multiplier that specifies the duration of each unit of time. For
example, if the timer returns times measured in thousands of seconds, the
time unit would be <code class="docutils literal notranslate">.001</code>.
Directly using the <a class="reference internal" href="#profile.Profile" title="profile.Profile"><code class="xref py py-class docutils literal notranslate">Profile</code></a> class allows formatting profile results
without writing the profile data to a file:
<div class="highlight-default notranslate"><div class="highlight"><pre>import cProfile, pstats, StringIO
pr = cProfile.Profile()
pr.enable()
# ... do something ...
pr.disable()
s = StringIO.StringIO()
sortby = &#39;cumulative&#39;
ps = pstats.Stats(pr, stream=s).sort_stats(sortby)
ps.print_stats()
print s.getvalue()
</pre></div>
</div>
<dl class="py method">
<dt class="sig sig-object py" id="profile.Profile.enable">
enable()<a class="headerlink" href="#profile.Profile.enable" title="Permalink to this definition">¶</a></dt>
<dd>Start collecting profiling data.
</dd></dl>

<dl class="py method">
<dt class="sig sig-object py" id="profile.Profile.disable">
disable()<a class="headerlink" href="#profile.Profile.disable" title="Permalink to this definition">¶</a></dt>
<dd>Stop collecting profiling data.
</dd></dl>

<dl class="py method">
<dt class="sig sig-object py" id="profile.Profile.create_stats">
create_stats()<a class="headerlink" href="#profile.Profile.create_stats" title="Permalink to this definition">¶</a></dt>
<dd>Stop collecting profiling data and record the results internally
as the current profile.
</dd></dl>

<dl class="py method">
<dt class="sig sig-object py" id="profile.Profile.print_stats">
print_stats(sort=- 1)<a class="headerlink" href="#profile.Profile.print_stats" title="Permalink to this definition">¶</a></dt>
<dd>Create a <a class="reference internal" href="#pstats.Stats" title="pstats.Stats"><code class="xref py py-class docutils literal notranslate">Stats</code></a> object based on the current
profile and print the results to stdout.
</dd></dl>

<dl class="py method">
<dt class="sig sig-object py" id="profile.Profile.dump_stats">
dump_stats(filename)<a class="headerlink" href="#profile.Profile.dump_stats" title="Permalink to this definition">¶</a></dt>
<dd>Write the results of the current profile to filename.
</dd></dl>

<dl class="py method">
<dt class="sig sig-object py" id="profile.Profile.run">
run(cmd)<a class="headerlink" href="#profile.Profile.run" title="Permalink to this definition">¶</a></dt>
<dd>Profile the cmd via <code class="xref py py-func docutils literal notranslate">exec()</code>.
</dd></dl>

<dl class="py method">
<dt class="sig sig-object py" id="profile.Profile.runctx">
runctx(cmd, globals, locals)<a class="headerlink" href="#profile.Profile.runctx" title="Permalink to this definition">¶</a></dt>
<dd>Profile the cmd via <code class="xref py py-func docutils literal notranslate">exec()</code> with the specified global and
local environment.
</dd></dl>

<dl class="py method">
<dt class="sig sig-object py" id="profile.Profile.runcall">
runcall(func, *args, **kwargs)<a class="headerlink" href="#profile.Profile.runcall" title="Permalink to this definition">¶</a></dt>
<dd>Profile <code class="docutils literal notranslate">func(*args, **kwargs)</code>
</dd></dl>

</dd></dl>

</section>
<section id="the-stats-class">
<h2>26.4.4. The <code class="xref py py-class docutils literal notranslate">Stats</code> Class<a class="headerlink" href="#the-stats-class" title="Permalink to this headline">¶</a></h2>
Analysis of the profiler data is done using the <a class="reference internal" href="#pstats.Stats" title="pstats.Stats"><code class="xref py py-class docutils literal notranslate">Stats</code></a> class.
<dl class="py class">
<dt class="sig sig-object py" id="pstats.Stats">
class pstats.Stats(*filenames or profile, stream=sys.stdout)<a class="headerlink" href="#pstats.Stats" title="Permalink to this definition">¶</a></dt>
<dd>This class constructor creates an instance of a “statistics object” from a
filename (or list of filenames) or from a <code class="xref py py-class docutils literal notranslate">Profile</code> instance. Output
will be printed to the stream specified by stream.
The file selected by the above constructor must have been created by the
corresponding version of <a class="reference internal" href="#module-profile" title="profile: Python source profiler."><code class="xref py py-mod docutils literal notranslate">profile</code></a> or <a class="reference internal" href="#module-cProfile" title="cProfile"><code class="xref py py-mod docutils literal notranslate">cProfile</code></a>. To be specific,
there is no file compatibility guaranteed with future versions of this
profiler, and there is no compatibility with files produced by other
profilers, or the same profiler run on a different operating system. If
several files are provided, all the statistics for identical functions will
be coalesced, so that an overall view of several processes can be considered
in a single report. If additional files need to be combined with data in an
existing <a class="reference internal" href="#pstats.Stats" title="pstats.Stats"><code class="xref py py-class docutils literal notranslate">Stats</code></a> object, the <a class="reference internal" href="#pstats.Stats.add" title="pstats.Stats.add"><code class="xref py py-meth docutils literal notranslate">add()</code></a> method
can be used.
Instead of reading the profile data from a file, a <code class="xref py py-class docutils literal notranslate">cProfile.Profile</code>
or <a class="reference internal" href="#profile.Profile" title="profile.Profile"><code class="xref py py-class docutils literal notranslate">profile.Profile</code></a> object can be used as the profile data source.
<a class="reference internal" href="#pstats.Stats" title="pstats.Stats"><code class="xref py py-class docutils literal notranslate">Stats</code></a> objects have the following methods:
<dl class="py method">
<dt class="sig sig-object py" id="pstats.Stats.strip_dirs">
strip_dirs()<a class="headerlink" href="#pstats.Stats.strip_dirs" title="Permalink to this definition">¶</a></dt>
<dd>This method for the <a class="reference internal" href="#pstats.Stats" title="pstats.Stats"><code class="xref py py-class docutils literal notranslate">Stats</code></a> class removes all leading path
information from file names. It is very useful in reducing the size of
the printout to fit within (close to) 80 columns. This method modifies
the object, and the stripped information is lost. After performing a
strip operation, the object is considered to have its entries in a
“random” order, as it was just after object initialization and loading.
If <a class="reference internal" href="#pstats.Stats.strip_dirs" title="pstats.Stats.strip_dirs"><code class="xref py py-meth docutils literal notranslate">strip_dirs()</code></a> causes two function names to be
indistinguishable (they are on the same line of the same filename, and
have the same function name), then the statistics for these two entries
are accumulated into a single entry.
</dd></dl>

<dl class="py method">
<dt class="sig sig-object py" id="pstats.Stats.add">
add(*filenames)<a class="headerlink" href="#pstats.Stats.add" title="Permalink to this definition">¶</a></dt>
<dd>This method of the <a class="reference internal" href="#pstats.Stats" title="pstats.Stats"><code class="xref py py-class docutils literal notranslate">Stats</code></a> class accumulates additional profiling
information into the current profiling object. Its arguments should refer
to filenames created by the corresponding version of <a class="reference internal" href="#profile.run" title="profile.run"><code class="xref py py-func docutils literal notranslate">profile.run()</code></a>
or <code class="xref py py-func docutils literal notranslate">cProfile.run()</code>. Statistics for identically named (re: file, line,
name) functions are automatically accumulated into single function
statistics.
</dd></dl>

<dl class="py method">
<dt class="sig sig-object py" id="pstats.Stats.dump_stats">
dump_stats(filename)<a class="headerlink" href="#pstats.Stats.dump_stats" title="Permalink to this definition">¶</a></dt>
<dd>Save the data loaded into the <a class="reference internal" href="#pstats.Stats" title="pstats.Stats"><code class="xref py py-class docutils literal notranslate">Stats</code></a> object to a file named
filename. The file is created if it does not exist, and is overwritten
if it already exists. This is equivalent to the method of the same name
on the <a class="reference internal" href="#profile.Profile" title="profile.Profile"><code class="xref py py-class docutils literal notranslate">profile.Profile</code></a> and <code class="xref py py-class docutils literal notranslate">cProfile.Profile</code> classes.
</dd></dl>

<div class="versionadded">
New in version 2.3.
</div>
<dl class="py method">
<dt class="sig sig-object py" id="pstats.Stats.sort_stats">
sort_stats(*keys)<a class="headerlink" href="#pstats.Stats.sort_stats" title="Permalink to this definition">¶</a></dt>
<dd>This method modifies the <a class="reference internal" href="#pstats.Stats" title="pstats.Stats"><code class="xref py py-class docutils literal notranslate">Stats</code></a> object by sorting it according to
the supplied criteria. The argument is typically a string identifying the
basis of a sort (example: <code class="docutils literal notranslate">'time'</code> or <code class="docutils literal notranslate">'name'</code>).
When more than one key is provided, then additional keys are used as
secondary criteria when there is equality in all keys selected before
them. For example, <code class="docutils literal notranslate">sort_stats('name', 'file')</code> will sort all the
entries according to their function name, and resolve all ties (identical
function names) by sorting by file name.
Abbreviations can be used for any key names, as long as the abbreviation
is unambiguous. The following are the keys currently defined:
<table class="docutils align-default">
<colgroup>
<col style="width: 45%" />
<col style="width: 55%" />
</colgroup>
<thead>
<tr class="row-odd"><th class="head">Valid Arg</th>
<th class="head">Meaning</th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><code class="docutils literal notranslate">'calls'</code></td>
<td>call count</td>
</tr>
<tr class="row-odd"><td><code class="docutils literal notranslate">'cumulative'</code></td>
<td>cumulative time</td>
</tr>
<tr class="row-even"><td><code class="docutils literal notranslate">'cumtime'</code></td>
<td>cumulative time</td>
</tr>
<tr class="row-odd"><td><code class="docutils literal notranslate">'file'</code></td>
<td>file name</td>
</tr>
<tr class="row-even"><td><code class="docutils literal notranslate">'filename'</code></td>
<td>file name</td>
</tr>
<tr class="row-odd"><td><code class="docutils literal notranslate">'module'</code></td>
<td>file name</td>
</tr>
<tr class="row-even"><td><code class="docutils literal notranslate">'ncalls'</code></td>
<td>call count</td>
</tr>
<tr class="row-odd"><td><code class="docutils literal notranslate">'pcalls'</code></td>
<td>primitive call count</td>
</tr>
<tr class="row-even"><td><code class="docutils literal notranslate">'line'</code></td>
<td>line number</td>
</tr>
<tr class="row-odd"><td><code class="docutils literal notranslate">'name'</code></td>
<td>function name</td>
</tr>
<tr class="row-even"><td><code class="docutils literal notranslate">'nfl'</code></td>
<td>name/file/line</td>
</tr>
<tr class="row-odd"><td><code class="docutils literal notranslate">'stdname'</code></td>
<td>standard name</td>
</tr>
<tr class="row-even"><td><code class="docutils literal notranslate">'time'</code></td>
<td>internal time</td>
</tr>
<tr class="row-odd"><td><code class="docutils literal notranslate">'tottime'</code></td>
<td>internal time</td>
</tr>
</tbody>
</table>
Note that all sorts on statistics are in descending order (placing most
time consuming items first), where as name, file, and line number searches
are in ascending order (alphabetical). The subtle distinction between
<code class="docutils literal notranslate">'nfl'</code> and <code class="docutils literal notranslate">'stdname'</code> is that the standard name is a sort of the
name as printed, which means that the embedded line numbers get compared
in an odd way. For example, lines 3, 20, and 40 would (if the file names
were the same) appear in the string order 20, 3 and 40. In contrast,
<code class="docutils literal notranslate">'nfl'</code> does a numeric compare of the line numbers. In fact,
<code class="docutils literal notranslate">sort_stats('nfl')</code> is the same as <code class="docutils literal notranslate">sort_stats('name', 'file',
'line')</code>.
For backward-compatibility reasons, the numeric arguments <code class="docutils literal notranslate">-1</code>, <code class="docutils literal notranslate">0</code>,
<code class="docutils literal notranslate">1</code>, and <code class="docutils literal notranslate">2</code> are permitted. They are interpreted as <code class="docutils literal notranslate">'stdname'</code>,
<code class="docutils literal notranslate">'calls'</code>, <code class="docutils literal notranslate">'time'</code>, and <code class="docutils literal notranslate">'cumulative'</code> respectively. If this old
style format (numeric) is used, only one sort key (the numeric key) will
be used, and additional arguments will be silently ignored.
</dd></dl>

<dl class="py method">
<dt class="sig sig-object py" id="pstats.Stats.reverse_order">
reverse_order()<a class="headerlink" href="#pstats.Stats.reverse_order" title="Permalink to this definition">¶</a></dt>
<dd>This method for the <a class="reference internal" href="#pstats.Stats" title="pstats.Stats"><code class="xref py py-class docutils literal notranslate">Stats</code></a> class reverses the ordering of the
basic list within the object. Note that by default ascending vs
descending order is properly selected based on the sort key of choice.
</dd></dl>

<dl class="py method">
<dt class="sig sig-object py" id="pstats.Stats.print_stats">
print_stats(*restrictions)<a class="headerlink" href="#pstats.Stats.print_stats" title="Permalink to this definition">¶</a></dt>
<dd>This method for the <a class="reference internal" href="#pstats.Stats" title="pstats.Stats"><code class="xref py py-class docutils literal notranslate">Stats</code></a> class prints out a report as described
in the <a class="reference internal" href="#profile.run" title="profile.run"><code class="xref py py-func docutils literal notranslate">profile.run()</code></a> definition.
The order of the printing is based on the last
<a class="reference internal" href="#pstats.Stats.sort_stats" title="pstats.Stats.sort_stats"><code class="xref py py-meth docutils literal notranslate">sort_stats()</code></a> operation done on the object (subject to
caveats in <a class="reference internal" href="#pstats.Stats.add" title="pstats.Stats.add"><code class="xref py py-meth docutils literal notranslate">add()</code></a> and
<a class="reference internal" href="#pstats.Stats.strip_dirs" title="pstats.Stats.strip_dirs"><code class="xref py py-meth docutils literal notranslate">strip_dirs()</code></a>).
The arguments provided (if any) can be used to limit the list down to the
significant entries. Initially, the list is taken to be the complete set
of profiled functions. Each restriction is either an integer (to select a
count of lines), or a decimal fraction between 0.0 and 1.0 inclusive (to
select a percentage of lines), or a regular expression (to pattern match
the standard name that is printed. If several restrictions are provided,
then they are applied sequentially. For example:
<div class="highlight-default notranslate"><div class="highlight"><pre>print_stats(.1, &#39;foo:&#39;)
</pre></div>
</div>
would first limit the printing to first 10% of list, and then only print
functions that were part of filename <code class="file docutils literal notranslate">.*foo:</code>. In contrast, the
command:
<div class="highlight-default notranslate"><div class="highlight"><pre>print_stats(&#39;foo:&#39;, .1)
</pre></div>
</div>
would limit the list to all functions having file names <code class="file docutils literal notranslate">.*foo:</code>,
and then proceed to only print the first 10% of them.
</dd></dl>

<dl class="py method">
<dt class="sig sig-object py" id="pstats.Stats.print_callers">
print_callers(*restrictions)<a class="headerlink" href="#pstats.Stats.print_callers" title="Permalink to this definition">¶</a></dt>
<dd>This method for the <a class="reference internal" href="#pstats.Stats" title="pstats.Stats"><code class="xref py py-class docutils literal notranslate">Stats</code></a> class prints a list of all functions
that called each function in the profiled database. The ordering is
identical to that provided by <a class="reference internal" href="#pstats.Stats.print_stats" title="pstats.Stats.print_stats"><code class="xref py py-meth docutils literal notranslate">print_stats()</code></a>, and the
definition of the restricting argument is also identical. Each caller is
reported on its own line. The format differs slightly depending on the
profiler that produced the stats:
<ul class="simple">
<li>With <a class="reference internal" href="#module-profile" title="profile: Python source profiler."><code class="xref py py-mod docutils literal notranslate">profile</code></a>, a number is shown in parentheses after each caller
to show how many times this specific call was made. For convenience, a
second non-parenthesized number repeats the cumulative time spent in the
function at the right.</li>
<li>With <a class="reference internal" href="#module-cProfile" title="cProfile"><code class="xref py py-mod docutils literal notranslate">cProfile</code></a>, each caller is preceded by three numbers: the
number of times this specific call was made, and the total and
cumulative times spent in the current function while it was invoked by
this specific caller.</li>
</ul>
</dd></dl>

<dl class="py method">
<dt class="sig sig-object py" id="pstats.Stats.print_callees">
print_callees(*restrictions)<a class="headerlink" href="#pstats.Stats.print_callees" title="Permalink to this definition">¶</a></dt>
<dd>This method for the <a class="reference internal" href="#pstats.Stats" title="pstats.Stats"><code class="xref py py-class docutils literal notranslate">Stats</code></a> class prints a list of all function
that were called by the indicated function. Aside from this reversal of
direction of calls (re: called vs was called by), the arguments and
ordering are identical to the <a class="reference internal" href="#pstats.Stats.print_callers" title="pstats.Stats.print_callers"><code class="xref py py-meth docutils literal notranslate">print_callers()</code></a> method.
</dd></dl>

</dd></dl>

</section>
<section id="what-is-deterministic-profiling">
<h2>26.4.5. What Is Deterministic Profiling?<a class="headerlink" href="#what-is-deterministic-profiling" title="Permalink to this headline">¶</a></h2>
Deterministic profiling is meant to reflect the fact that all function
call, function return, and exception events are monitored, and precise
timings are made for the intervals between these events (during which time the
user’s code is executing). In contrast, statistical profiling (which is
not done by this module) randomly samples the effective instruction pointer, and
deduces where time is being spent. The latter technique traditionally involves
less overhead (as the code does not need to be instrumented), but provides only
relative indications of where time is being spent.
In Python, since there is an interpreter active during execution, the presence
of instrumented code is not required to do deterministic profiling. Python
automatically provides a hook (optional callback) for each event. In
addition, the interpreted nature of Python tends to add so much overhead to
execution, that deterministic profiling tends to only add small processing
overhead in typical applications. The result is that deterministic profiling is
not that expensive, yet provides extensive run time statistics about the
execution of a Python program.
Call count statistics can be used to identify bugs in code (surprising counts),
and to identify possible inline-expansion points (high call counts). Internal
time statistics can be used to identify “hot loops” that should be carefully
optimized. Cumulative time statistics should be used to identify high level
errors in the selection of algorithms. Note that the unusual handling of
cumulative times in this profiler allows statistics for recursive
implementations of algorithms to be directly compared to iterative
implementations.
</section>
<section id="limitations">
<h2>26.4.6. Limitations<a class="headerlink" href="#limitations" title="Permalink to this headline">¶</a></h2>
One limitation has to do with accuracy of timing information. There is a
fundamental problem with deterministic profilers involving accuracy. The most
obvious restriction is that the underlying “clock” is only ticking at a rate
(typically) of about .001 seconds. Hence no measurements will be more accurate
than the underlying clock. If enough measurements are taken, then the “error”
will tend to average out. Unfortunately, removing this first error induces a
second source of error.
The second problem is that it “takes a while” from when an event is dispatched
until the profiler’s call to get the time actually gets the state of the
clock. Similarly, there is a certain lag when exiting the profiler event
handler from the time that the clock’s value was obtained (and then squirreled
away), until the user’s code is once again executing. As a result, functions
that are called many times, or call many functions, will typically accumulate
this error. The error that accumulates in this fashion is typically less than
the accuracy of the clock (less than one clock tick), but it can accumulate
and become very significant.
The problem is more important with <a class="reference internal" href="#module-profile" title="profile: Python source profiler."><code class="xref py py-mod docutils literal notranslate">profile</code></a> than with the lower-overhead
<a class="reference internal" href="#module-cProfile" title="cProfile"><code class="xref py py-mod docutils literal notranslate">cProfile</code></a>. For this reason, <a class="reference internal" href="#module-profile" title="profile: Python source profiler."><code class="xref py py-mod docutils literal notranslate">profile</code></a> provides a means of
calibrating itself for a given platform so that this error can be
probabilistically (on the average) removed. After the profiler is calibrated, it
will be more accurate (in a least square sense), but it will sometimes produce
negative numbers (when call counts are exceptionally low, and the gods of
probability work against you :-). ) Do not be alarmed by negative numbers in
the profile. They should only appear if you have calibrated your profiler,
and the results are actually better than without calibration.
</section>
<section id="calibration">
<h2>26.4.7. Calibration<a class="headerlink" href="#calibration" title="Permalink to this headline">¶</a></h2>
The profiler of the <a class="reference internal" href="#module-profile" title="profile: Python source profiler."><code class="xref py py-mod docutils literal notranslate">profile</code></a> module subtracts a constant from each event
handling time to compensate for the overhead of calling the time function, and
socking away the results. By default, the constant is 0. The following
procedure can be used to obtain a better constant for a given platform (see
<a class="reference internal" href="#profile-limitations">Limitations</a>).
<div class="highlight-default notranslate"><div class="highlight"><pre>import profile
pr = profile.Profile()
for i in range(5):
 print pr.calibrate(10000)
</pre></div>
</div>
The method executes the number of Python calls given by the argument, directly
and again under the profiler, measuring the time for both. It then computes the
hidden overhead per profiler event, and returns that as a float. For example,
on a 1.8Ghz Intel Core i5 running Mac OS X, and using Python’s time.clock() as
the timer, the magical number is about 4.04e-6.
The object of this exercise is to get a fairly consistent result. If your
computer is very fast, or your timer function has poor resolution, you might
have to pass 100000, or even 1000000, to get consistent results.
When you have a consistent answer, there are three ways you can use it: <a class="footnote-reference brackets" href="#id2" id="id1">1</a>
<div class="highlight-default notranslate"><div class="highlight"><pre>import profile

# 1. Apply computed bias to all Profile instances created hereafter.
profile.Profile.bias = your_computed_bias

# 2. Apply computed bias to a specific Profile instance.
pr = profile.Profile()
pr.bias = your_computed_bias

# 3. Specify computed bias in instance constructor.
pr = profile.Profile(bias=your_computed_bias)
</pre></div>
</div>
If you have a choice, you are better off choosing a smaller constant, and then
your results will “less often” show up as negative in profile statistics.
</section>
<section id="using-a-custom-timer">
<h2>26.4.8. Using a custom timer<a class="headerlink" href="#using-a-custom-timer" title="Permalink to this headline">¶</a></h2>
If you want to change how current time is determined (for example, to force use
of wall-clock time or elapsed process time), pass the timing function you want
to the <code class="xref py py-class docutils literal notranslate">Profile</code> class constructor:
<div class="highlight-default notranslate"><div class="highlight"><pre>pr = profile.Profile(your_time_func)
</pre></div>
</div>
The resulting profiler will then call <code class="docutils literal notranslate">your_time_func</code>. Depending on whether
you are using <a class="reference internal" href="#profile.Profile" title="profile.Profile"><code class="xref py py-class docutils literal notranslate">profile.Profile</code></a> or <code class="xref py py-class docutils literal notranslate">cProfile.Profile</code>,
<code class="docutils literal notranslate">your_time_func</code>’s return value will be interpreted differently:
<dl>
<dt><a class="reference internal" href="#profile.Profile" title="profile.Profile"><code class="xref py py-class docutils literal notranslate">profile.Profile</code></a></dt><dd><code class="docutils literal notranslate">your_time_func</code> should return a single number, or a list of numbers whose
sum is the current time (like what <a class="reference internal" href="os.html#os.times" title="os.times"><code class="xref py py-func docutils literal notranslate">os.times()</code></a> returns). If the
function returns a single time number, or the list of returned numbers has
length 2, then you will get an especially fast version of the dispatch
routine.
Be warned that you should calibrate the profiler class for the timer function
that you choose (see <a class="reference internal" href="#profile-calibration">Calibration</a>). For most machines, a timer
that returns a lone integer value will provide the best results in terms of
low overhead during profiling. (<a class="reference internal" href="os.html#os.times" title="os.times"><code class="xref py py-func docutils literal notranslate">os.times()</code></a> is pretty bad, as it
returns a tuple of floating point values). If you want to substitute a
better timer in the cleanest fashion, derive a class and hardwire a
replacement dispatch method that best handles your timer call, along with the
appropriate calibration constant.
</dd>
<dt><code class="xref py py-class docutils literal notranslate">cProfile.Profile</code></dt><dd><code class="docutils literal notranslate">your_time_func</code> should return a single number. If it returns integers,
you can also invoke the class constructor with a second argument specifying
the real duration of one unit of time. For example, if
<code class="docutils literal notranslate">your_integer_time_func</code> returns times measured in thousands of seconds,
you would construct the <code class="xref py py-class docutils literal notranslate">Profile</code> instance as follows:
<div class="highlight-default notranslate"><div class="highlight"><pre>pr = cProfile.Profile(your_integer_time_func, 0.001)
</pre></div>
</div>
As the <code class="xref py py-class docutils literal notranslate">cProfile.Profile</code> class cannot be calibrated, custom timer
functions should be used with care and should be as fast as possible. For
the best results with a custom timer, it might be necessary to hard-code it
in the C source of the internal <code class="xref py py-mod docutils literal notranslate">_lsprof</code> module.
</dd>
</dl>
Footnotes
<dl class="footnote brackets">
<dt class="label" id="id2"><a class="fn-backref" href="#id1">1</a></dt>
<dd>Prior to Python 2.2, it was necessary to edit the profiler source code to
embed the bias as a literal number. You still can, but that method is no longer
described, because no longer needed.
</dd>
</dl>
</section>
</section>

<div class="clearer"></div>
 </div>
 </div>
 </div>
 <div class="sphinxsidebar" role="navigation" aria-label="main navigation">
 <div class="sphinxsidebarwrapper">
 <h3><a href="../contents.html">Table of Contents</a></h3>
 <ul>
<li><a class="reference internal" href="#">26.4. The Python Profilers</a><ul>
<li><a class="reference internal" href="#introduction-to-the-profilers">26.4.1. Introduction to the profilers</a></li>
<li><a class="reference internal" href="#instant-user-s-manual">26.4.2. Instant User’s Manual</a></li>
<li><a class="reference internal" href="#module-cProfile">26.4.3. <code class="xref py py-mod docutils literal notranslate">profile</code> and <code class="xref py py-mod docutils literal notranslate">cProfile</code> Module Reference</a></li>
<li><a class="reference internal" href="#the-stats-class">26.4.4. The <code class="xref py py-class docutils literal notranslate">Stats</code> Class</a></li>
<li><a class="reference internal" href="#what-is-deterministic-profiling">26.4.5. What Is Deterministic Profiling?</a></li>
<li><a class="reference internal" href="#limitations">26.4.6. Limitations</a></li>
<li><a class="reference internal" href="#calibration">26.4.7. Calibration</a></li>
<li><a class="reference internal" href="#using-a-custom-timer">26.4.8. Using a custom timer</a></li>
</ul>
</li>
</ul>

<h4>Previous topic</h4>
 <a href="pdb.html"
 title="previous chapter">26.2. <code class="xref py py-mod docutils literal notranslate">pdb</code> — The Python Debugger</a>
 <h4>Next topic</h4>
 <a href="hotshot.html"
 title="next chapter">26.5. <code class="xref py py-mod docutils literal notranslate">hotshot</code> — High performance logging profiler</a>
 <div role="note" aria-label="source link">
 <h3>This Page</h3>
 <ul class="this-page-menu">
 <li><a href="../_sources/library/profile.rst.txt"
 rel="nofollow">Show Source</a></li>
 </ul>
 </div>
<div id="searchbox" style="display: none" role="search">
 <h3 id="searchlabel">Quick search</h3>
 <div class="searchformwrapper">
 <form class="search" action="../search.html" method="get">
 <input type="text" name="q" aria-labelledby="searchlabel" autocomplete="off" autocorrect="off" autocapitalize="off" spellcheck="false"/>
 <input type="submit" value="Go" />
 </form>
 </div>
</div>
<script>$('#searchbox').show(0);</script>
 </div>
 </div>
 <div class="clearer"></div>
 </div> 
 <div class="related" role="navigation" aria-label="related navigation">
 <h3>Navigation</h3>
 <ul>
 <li class="right" style="margin-right: 10px">
 <a href="../genindex.html" title="General Index"
 >index</a></li>
 <li class="right" >
 <a href="../py-modindex.html" title="Python Module Index"
 >modules</a> |</li>
 <li class="right" >
 <a href="hotshot.html" title="26.5. hotshot — High performance logging profiler"
 >next</a> |</li>
 <li class="right" >
 <a href="pdb.html" title="26.2. pdb — The Python Debugger"
 >previous</a> |</li>
 <li><img src="../_static/py.png" alt=""
 style="vertical-align: middle; margin-top: -1px"/></li>
 <li><a href="https://www.python.org/">Python</a> &#187;</li>
 <li>
 <a href="../index.html">Python 2.7.18 documentation</a> &#187;
 </li>

<li class="nav-item nav-item-1"><a href="index.html" >The Python Standard Library</a> &#187;</li>
 <li class="nav-item nav-item-2"><a href="debug.html" >26. Debugging and Profiling</a> &#187;</li>
 <li class="nav-item nav-item-this"><a href="">26.4. The Python Profilers</a></li> 
 </ul>
 </div> 
 <div class="footer">
 &copy; <a href="../copyright.html">Copyright</a> 1990-2024, Python Software Foundation.
 
 The Python Software Foundation is a non-profit corporation.
 <a href="https://www.python.org/psf/donations/">Please donate.</a>
 
 Last updated on November 21, 2024.
 <a href="../bugs.html">Found a bug</a>?
 
 Created using <a href="http://sphinx.pocoo.org/">Sphinx</a> 4.3.2.
 </div>

</body>
</html>