<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<title>FAQ — PyTables 3.0.0 documentation</title>
<link rel="stylesheet" href="_static/cloud.css" type="text/css" />
<link rel="stylesheet" href="_static/pygments.css" type="text/css" />
<link rel="stylesheet" href="#" type="text/css" />
<script type="text/javascript">
var DOCUMENTATION_OPTIONS = {
URL_ROOT: '',
VERSION: '3.0.0',
COLLAPSE_INDEX: false,
FILE_SUFFIX: '.html',
HAS_SOURCE: true
};
</script>
<script type="text/javascript" src="_static/jquery.js"></script>
<script type="text/javascript" src="_static/underscore.js"></script>
<script type="text/javascript" src="_static/doctools.js"></script>
<script type="text/javascript" src="_static/jquery.cookie.js"></script>
<script type="text/javascript" src="_static/toggle_sections.js"></script>
<script type="text/javascript" src="_static/toggle_sidebar.js"></script>
<link rel="shortcut icon" href="_static/favicon.ico"/>
<link rel="top" title="PyTables 3.0.0 documentation" href="index.html" />
<link rel="next" title="Other Material" href="other_material.html" />
<link rel="prev" title="In-memory HDF5 files" href="cookbook/inmemory_hdf5_files.html" />
</head>
<body>
<div class="relbar-top">
<div class="related">
<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="np-modindex.html" title="Python Module Index"
>modules</a> </li>
<li class="right" >
<a href="other_material.html" title="Other Material"
accesskey="N">next</a> </li>
<li class="right" >
<a href="cookbook/inmemory_hdf5_files.html" title="In-memory HDF5 files"
accesskey="P">previous</a> </li>
<li><a href="index.html">PyTables 3.0.0 documentation</a> »</li>
</ul>
</div>
</div>
<div class="document">
<div class="documentwrapper">
<div class="bodywrapper">
<div class="body">
<div class="section" id="faq">
<h1>FAQ<a class="headerlink" href="#faq" title="Permalink to this headline">¶</a></h1>
<div class="section" id="general-questions">
<h2>General questions<a class="headerlink" href="#general-questions" title="Permalink to this headline">¶</a></h2>
<div class="section" id="what-is-pytables">
<h3>What is PyTables?<a class="headerlink" href="#what-is-pytables" title="Permalink to this headline">¶</a></h3>
<p>PyTables is a package for managing hierarchical datasets designed to
efficiently cope with extremely large amounts of data.</p>
<p>It is built on top of the <a class="reference external" href="http://www.hdfgroup.org/HDF5">HDF5</a> <a class="footnote-reference" href="#id2" id="id3">[1]</a> library, the <a class="reference external" href="http://www.python.org">Python language</a> <a class="footnote-reference" href="#id18" id="id19">[2]</a> and the
<a class="reference external" href="http://www.numpy.org">NumPy</a> <a class="footnote-reference" href="#id20" id="id21">[3]</a> package.
It features an object-oriented interface that, combined with C extensions
for the performance-critical parts of the code, makes it a fast yet
extremely easy-to-use tool for interactively storing and retrieving very
large amounts of data.</p>
</div>
<div class="section" id="what-are-pytables-licensing-terms">
<h3>What are PyTables’ licensing terms?<a class="headerlink" href="#what-are-pytables-licensing-terms" title="Permalink to this headline">¶</a></h3>
<p>PyTables is free for both commercial and non-commercial use, under the terms
of the BSD license.</p>
</div>
<div class="section" id="i-m-having-problems-how-can-i-get-support">
<h3>I’m having problems. How can I get support?<a class="headerlink" href="#i-m-having-problems-how-can-i-get-support" title="Permalink to this headline">¶</a></h3>
<p>The most common and efficient way is to subscribe (remember you <em>need</em> to
subscribe prior to send messages) to the PyTables <a class="reference external" href="https://lists.sourceforge.net/lists/listinfo/pytables-users">users mailing list</a> <a class="footnote-reference" href="#id31" id="id32">[4]</a>, and
send there a brief description of your issue and, if possible, a short script
that can reproduce it.
Hopefully, someone on the list will be able to help you.
It is also a good idea to check out the <a class="reference external" href="http://sourceforge.net/mailarchive/forum.php?forum_id=13760">archives of the user’s list</a> <a class="footnote-reference" href="#id33" id="id34">[5]</a> (you may
want to check the <a class="reference external" href="http://www.mail-archive.com/pytables-users@lists.sourceforge.net/">Gmane archives</a> <a class="footnote-reference" href="#id35" id="id36">[6]</a> instead) so as to see if the answer to your
question has already been dealed with.</p>
</div>
<div class="section" id="why-hdf5">
<h3>Why HDF5?<a class="headerlink" href="#why-hdf5" title="Permalink to this headline">¶</a></h3>
<p><a class="reference external" href="http://www.hdfgroup.org/HDF5">HDF5</a> <a class="footnote-reference" href="#id2" id="id4">[1]</a> is the underlying C library and file format that enables PyTables to
efficiently deal with the data. It has been chosen for the following reasons:</p>
<ul class="simple">
<li>Designed to efficiently manage very large datasets.</li>
<li>Lets you organize datasets hierarchically.</li>
<li>Very flexible and well tested in scientific environments.</li>
<li>Good maintenance and improvement rate.</li>
<li>Technical excellence (<a class="reference external" href="http://www.hdfgroup.org/HDF5/RD100-2002/">R&D 100 Award</a> <a class="footnote-reference" href="#id37" id="id38">[7]</a>).</li>
<li><strong>It’s Open Source software</strong></li>
</ul>
</div>
<div class="section" id="why-python">
<h3>Why Python?<a class="headerlink" href="#why-python" title="Permalink to this headline">¶</a></h3>
<ol class="arabic">
<li><p class="first">Python is interactive.</p>
<p>People familiar with data processing understand how powerful command line
interfaces are for exploring mathematical relationships and scientific data
sets. Python provides an interactive environment with the added benefit of
a full featured programming language behind it.</p>
</li>
<li><p class="first">Python is productive for beginners and experts alike.</p>
<p>PyTables is targeted at engineers, scientists, system analysts, financial
analysts, and others who consider programming a necessary evil. Any time
spent learning a language or tracking down bugs is time spent not solving
their real problem. Python has a short learning curve and most people can
do real and useful work with it in a day of learning. Its clean syntax and
interactive nature facilitate this.</p>
</li>
<li><p class="first">Python is data-handling friendly.</p>
<p>Python comes with nice idioms that make the access to data much easier:
general slicing (i.e. <tt class="docutils literal"><span class="pre">data[start:stop:step]</span></tt>), list comprehensions,
iterators, generators ... are constructs that make the interaction with your
data very easy.</p>
</li>
</ol>
</div>
<div class="section" id="why-numpy">
<h3>Why NumPy?<a class="headerlink" href="#why-numpy" title="Permalink to this headline">¶</a></h3>
<p><a class="reference external" href="http://www.numpy.org">NumPy</a> <a class="footnote-reference" href="#id20" id="id22">[3]</a> is a Python package to efficiently deal with large datasets
<strong>in-memory</strong>, providing containers for homogeneous data, heterogeneous data,
and string arrays.
PyTables uses these NumPy containers as <em>in-memory buffers</em> to push the I/O
bandwith towards the platform limits.</p>
</div>
</div>
<div class="section" id="where-can-pytables-be-applied">
<h2>Where can PyTables be applied?<a class="headerlink" href="#where-can-pytables-be-applied" title="Permalink to this headline">¶</a></h2>
<p>In all the scenarios where one needs to deal with large datasets:</p>
<ul class="simple">
<li>Industrial applications<ul>
<li>Data acquisition in real time</li>
<li>Quality control</li>
<li>Fast data processing</li>
</ul>
</li>
<li>Scientific applications<ul>
<li>Meteorology, oceanography</li>
<li>Numerical simulations</li>
<li>Medicine (biological sensors, general data gathering & processing)</li>
</ul>
</li>
<li>Information systems<ul>
<li>System log monitoring & consolidation</li>
<li>Tracing of routing data</li>
<li>Alert systems in security</li>
</ul>
</li>
</ul>
<div class="section" id="is-pytables-safe">
<h3>Is PyTables safe?<a class="headerlink" href="#is-pytables-safe" title="Permalink to this headline">¶</a></h3>
<p>Well, first of all, let me state that PyTables does not support transactional
features yet (we don’t even know if we will ever be motivated to implement
this!), so there is always the risk that you can lose your data in case of an
unexpected event while writing (like a power outage, system shutdowns ...).
Having said that, if your typical scenarios are <em>write once, read many</em>, then
the use of PyTables is perfectly safe, even for dealing extremely large amounts
of data.</p>
</div>
<div class="section" id="can-pytables-be-used-in-concurrent-access-scenarios">
<h3>Can PyTables be used in concurrent access scenarios?<a class="headerlink" href="#can-pytables-be-used-in-concurrent-access-scenarios" title="Permalink to this headline">¶</a></h3>
<p>It depends. Concurrent reads are no problem at all. However, whenever a process
(or thread) is trying to write, then problems will start to appear. First,
PyTables doesn’t support locking at any level, so several process writing
concurrently to the same PyTables file will probably end up corrupting it, so
don’t do this! Even having only one process writing and the others reading is
a hairy thing, because the reading processes might be reading incomplete data
from a concurrent data writing operation.</p>
<p>The solution would be to lock the file while writing and unlock it after a
flush over the file has been performed. Also, in order to avoid cache (<a class="reference external" href="http://www.hdfgroup.org/HDF5">HDF5</a> <a class="footnote-reference" href="#id2" id="id5">[1]</a>,
PyTables) problems with read apps, you would need to re-open your files
whenever you are going to issue a read operation. If a re-opening operation is
unacceptable in terms of speed, you may want to do all your I/O operations in
one single process (or thread) and communicate the results via sockets,
<tt class="xref py py-class docutils literal"><span class="pre">Queue.Queue</span></tt> objects (in case of using threads), or whatever, with the
client process/thread.</p>
<p>The examples directory contains two scripts demonstrating methods of accessing a
PyTables file from multiple processes.</p>
<p>The first, <em>multiprocess_access_queues.py</em>, uses a
<tt class="xref py py-class docutils literal"><span class="pre">multiprocessing.Queue</span></tt> object to transfer read and write requests from
multiple <em>DataProcessor</em> processes to a single process responsible for all
access to the PyTables file. The results of read requests are then transferred
back to the originating processes using other <tt class="xref py py-class docutils literal"><span class="pre">Queue</span></tt> objects.</p>
<p>The second example script, <em>multiprocess_access_benchmarks.py</em>, demonstrates
and benchmarks four methods of transferring PyTables array data between
processes. The four methods are:</p>
<blockquote>
<div><ul class="simple">
<li>Using <tt class="xref py py-class docutils literal"><span class="pre">multiprocessing.Pipe</span></tt> from the Python standard library.</li>
<li>Using a memory mapped file that is shared between two processes. The NumPy
array associated with the file is passed as the <em>out</em> argument to the
<a class="reference internal" href="usersguide/libref/homogenous_storage.html#tables.Array.read" title="tables.Array.read"><tt class="xref py py-meth docutils literal"><span class="pre">tables.Array.read()</span></tt></a> method.</li>
<li>Using a Unix domain socket. Note that this example uses the ‘abstract
namespace’ and will only work under Linux.</li>
<li>Using an IPv4 socket.</li>
</ul>
</div></blockquote>
</div>
<div class="section" id="what-kind-of-containers-does-pytables-implement">
<h3>What kind of containers does PyTables implement?<a class="headerlink" href="#what-kind-of-containers-does-pytables-implement" title="Permalink to this headline">¶</a></h3>
<p>PyTables does support a series of data containers that address specific needs
of the user. Below is a brief description of them:</p>
<table class="docutils field-list" frame="void" rules="none">
<col class="field-name" />
<col class="field-body" />
<tbody valign="top">
<tr class="field-odd field"><th class="field-name">Table:</th><td class="field-body">Lets you deal with heterogeneous datasets. Allows compression. Enlargeable.
Supports nested types. Good performance for read/writing data.</td>
</tr>
<tr class="field-even field"><th class="field-name">Array:</th><td class="field-body">Provides quick and dirty array handling. Not compression allowed.
Not enlargeable. Can be used only with relatively small datasets (i.e.
those that fit in memory). It provides the fastest I/O speed.</td>
</tr>
<tr class="field-odd field"><th class="field-name">CArray:</th><td class="field-body">Provides compressed array support. Not enlargeable. Good speed when
reading/writing.</td>
</tr>
<tr class="field-even field"><th class="field-name">EArray:</th><td class="field-body">Most general array support. Compressible and enlargeable. It is pretty
fast at extending, and very good at reading.</td>
</tr>
<tr class="field-odd field"><th class="field-name">VLArray:</th><td class="field-body">Supports collections of homogeneous data with a variable number of entries.
Compressible and enlargeable. I/O is not very fast.</td>
</tr>
<tr class="field-even field"><th class="field-name">Group:</th><td class="field-body">The structural component.</td>
</tr>
</tbody>
</table>
<p>Please refer to the documentation for more specific information.</p>
</div>
<div class="section" id="cool-i-d-like-to-see-some-examples-of-use">
<h3>Cool! I’d like to see some examples of use.<a class="headerlink" href="#cool-i-d-like-to-see-some-examples-of-use" title="Permalink to this headline">¶</a></h3>
<p>Sure. Go to the HowToUse section to find simple examples that will help you
getting started.</p>
</div>
<div class="section" id="can-you-show-me-some-screenshots">
<h3>Can you show me some screenshots?<a class="headerlink" href="#can-you-show-me-some-screenshots" title="Permalink to this headline">¶</a></h3>
<p>Well, PyTables is not a graphical library by itself. However, you may want to
check out <a class="reference external" href="http://vitables.org">ViTables</a> <a class="footnote-reference" href="#id39" id="id40">[8]</a>, a GUI tool to browse and edit PyTables & <a class="reference external" href="http://www.hdfgroup.org/HDF5">HDF5</a> <a class="footnote-reference" href="#id2" id="id6">[1]</a> files.</p>
</div>
<div class="section" id="is-pytables-a-replacement-for-a-relational-database">
<h3>Is PyTables a replacement for a relational database?<a class="headerlink" href="#is-pytables-a-replacement-for-a-relational-database" title="Permalink to this headline">¶</a></h3>
<p>No, by no means. PyTables lacks many features that are standard in most
relational databases. In particular, it does not have support for
relationships (beyond the hierarchical one, of course) between datasets and it
does not have transactional features. PyTables is more focused on speed and
dealing with really large datasets, than implementing the above features. In
that sense, PyTables can be best viewed as a <em>teammate</em> of a relational
database.</p>
<p>For example, if you have very large tables in your existing relational
database, they will take lots of space on disk, potentially reducing the
performance of the relational engine. In such a case, you can move those huge
tables out of your existing relational database to PyTables, and let your
relational engine do what it does best (i.e. manage relatively small or medium
datasets with potentially complex relationships), and use PyTables for what it
has been designed for (i.e. manage large amounts of data which are loosely
related).</p>
</div>
<div class="section" id="how-can-pytables-be-fast-if-it-is-written-in-an-interpreted-language-like-python">
<h3>How can PyTables be fast if it is written in an interpreted language like Python?<a class="headerlink" href="#how-can-pytables-be-fast-if-it-is-written-in-an-interpreted-language-like-python" title="Permalink to this headline">¶</a></h3>
<p>Actually, all of the critical I/O code in PyTables is a thin layer of code on
top of <a class="reference external" href="http://www.hdfgroup.org/HDF5">HDF5</a> <a class="footnote-reference" href="#id2" id="id7">[1]</a>, which is a very efficient C library. <a class="reference external" href="http://www.cython.org">Cython</a> <a class="footnote-reference" href="#id41" id="id42">[9]</a> is used as the
<em>glue</em> language to generate “wrappers” around HDF5 calls so that they can be
used in Python. Also, the use of an efficient numerical package such as <a class="reference external" href="http://www.numpy.org">NumPy</a> <a class="footnote-reference" href="#id20" id="id23">[3]</a>
makes the most costly operations effectively run at C speed. Finally,
time-critical loops are usually implemented in <a class="reference external" href="http://www.cython.org">Cython</a> <a class="footnote-reference" href="#id41" id="id43">[9]</a> (which, if used
properly, allows to generate code that runs at almost pure C speeds).</p>
</div>
<div class="section" id="if-it-is-designed-to-deal-with-very-large-datasets-then-pytables-should-consume-a-lot-of-memory-shouldn-t-it">
<h3>If it is designed to deal with very large datasets, then PyTables should consume a lot of memory, shouldn’t it?<a class="headerlink" href="#if-it-is-designed-to-deal-with-very-large-datasets-then-pytables-should-consume-a-lot-of-memory-shouldn-t-it" title="Permalink to this headline">¶</a></h3>
<p>Well, you already know that PyTables sits on top of HDF5, Python and <a class="reference external" href="http://www.numpy.org">NumPy</a> <a class="footnote-reference" href="#id20" id="id24">[3]</a>,
and if we add its own logic (~7500 lines of code in Python, ~3000 in Cython and
~4000 in C), then we should conclude that PyTables isn’t effectively a paradigm
of lightness.</p>
<p>Having said that, PyTables (as <a class="reference external" href="http://www.hdfgroup.org/HDF5">HDF5</a> <a class="footnote-reference" href="#id2" id="id8">[1]</a> itself) tries very hard to optimize the
memory consumption by implementing a series of features like dynamic
determination of buffer sizes, <em>Least Recently Used</em> cache for keeping unused
nodes out of memory, and extensive use of compact <a class="reference external" href="http://www.numpy.org">NumPy</a> <a class="footnote-reference" href="#id20" id="id25">[3]</a> data containers.
Moreover, PyTables is in a relatively mature state and most memory leaks have
been already addressed and fixed.</p>
<p>Just to give you an idea of what you can expect, a PyTables program can deal
with a table with around 30 columns and 1 million entries using as low as 13 MB
of memory (on a 32-bit platform). All in all, it is not that much, is it?.</p>
</div>
<div class="section" id="why-was-pytables-born">
<h3>Why was PyTables born?<a class="headerlink" href="#why-was-pytables-born" title="Permalink to this headline">¶</a></h3>
<p>Because, back in August 2002, one of its authors (<a class="reference external" href="http://www.pytables.org/moin/FrancescAlted">Francesc Alted</a> <a class="footnote-reference" href="#id44" id="id45">[10]</a>) had a need
to save lots of hierarchical data in an efficient way for later post-processing
it. After trying out several approaches, he found that they presented distinct
inconveniences. For example, working with file sizes larger than, say, 100 MB,
was rather painful with ZODB (it took lots of memory with the version available
by that time).</p>
<p>The <a class="reference external" href="http://www.unidata.ucar.edu/software/netcdf">netCDF3</a> <a class="footnote-reference" href="#id46" id="id47">[11]</a> interface provided by <a class="reference external" href="http://dirac.cnrs-orleans.fr/plone/software/scientificpython">Scientific Python</a> <a class="footnote-reference" href="#id50" id="id51">[12]</a> was great, but it did
not allow to structure the hierarchically; besides, <a class="reference external" href="http://www.unidata.ucar.edu/software/netcdf">netCDF3</a> <a class="footnote-reference" href="#id46" id="id48">[11]</a> only supports
homogeneous datasets, not heterogeneous ones (i.e. tables). (As an aside,
<a class="reference external" href="http://www.unidata.ucar.edu/software/netcdf">netCDF4</a> <a class="footnote-reference" href="#id46" id="id52">[11]</a> overcomes many of the limitations of <a class="reference external" href="http://www.unidata.ucar.edu/software/netcdf">netCDF3</a> <a class="footnote-reference" href="#id46" id="id49">[11]</a>, although curiously
enough, it is based on top of <a class="reference external" href="http://www.hdfgroup.org/HDF5">HDF5</a> <a class="footnote-reference" href="#id2" id="id9">[1]</a>, the library chosen as the base for
PyTables from the very beginning.)</p>
<p>So, he decided to give <a class="reference external" href="http://www.hdfgroup.org/HDF5">HDF5</a> <a class="footnote-reference" href="#id2" id="id10">[1]</a> a try, start doing his own wrappings to it and
voilà, this is how the first public release of PyTables (0.1) saw the light in
October 2002, three months after his itch started to eat him ;-).</p>
</div>
<div class="section" id="does-pytables-have-a-client-server-interface">
<h3>Does PyTables have a client-server interface?<a class="headerlink" href="#does-pytables-have-a-client-server-interface" title="Permalink to this headline">¶</a></h3>
<p>Not by itself, but you may be interested in using PyTables through <a class="reference external" href="http://www.pydap.org">pydap</a> <a class="footnote-reference" href="#id53" id="id54">[13]</a>, a
Python implementation of the <a class="reference external" href="http://opendap.org">OPeNDAP</a> <a class="footnote-reference" href="#id56" id="id57">[14]</a> protocol. Have a look at the <cite>PyTables
plugin</cite> of <a class="reference external" href="http://www.pydap.org">pydap</a> <a class="footnote-reference" href="#id53" id="id55">[13]</a>.</p>
</div>
<div class="section" id="how-does-pytables-compare-with-the-h5py-project">
<h3>How does PyTables compare with the h5py project?<a class="headerlink" href="#how-does-pytables-compare-with-the-h5py-project" title="Permalink to this headline">¶</a></h3>
<p>Well, they are similar in that both packages are Python interfaces to the <a class="reference external" href="http://www.hdfgroup.org/HDF5">HDF5</a> <a class="footnote-reference" href="#id2" id="id11">[1]</a>
library, but there are some important differences to be noted. <a class="reference external" href="http://www.h5py.org">h5py</a> <a class="footnote-reference" href="#id60" id="id61">[16]</a> is an
attempt to map the <a class="reference external" href="http://www.hdfgroup.org/HDF5">HDF5</a> <a class="footnote-reference" href="#id2" id="id12">[1]</a> feature set to <a class="reference external" href="http://www.numpy.org">NumPy</a> <a class="footnote-reference" href="#id20" id="id26">[3]</a> as closely as possible. In
addition, it also provides access to nearly all of the <a class="reference external" href="http://www.hdfgroup.org/HDF5">HDF5</a> <a class="footnote-reference" href="#id2" id="id13">[1]</a> C API.</p>
<p>Instead, PyTables builds up an additional abstraction layer on top of <a class="reference external" href="http://www.hdfgroup.org/HDF5">HDF5</a> <a class="footnote-reference" href="#id2" id="id14">[1]</a> and
<a class="reference external" href="http://www.numpy.org">NumPy</a> <a class="footnote-reference" href="#id20" id="id27">[3]</a> where it implements things like an enhanced type system, an <a class="reference internal" href="usersguide/optimization.html#searchoptim"><em>engine
for enabling complex queries</em></a>, an <a class="reference external" href="http://www.pytables.org/moin/ComputingKernel">efficient computational
kernel</a> <a class="footnote-reference" href="#id62" id="id63">[17]</a>, <a class="reference external" href="http://www.pytables.org/moin/PyTablesPro">advanced indexing capabilities</a> <a class="footnote-reference" href="#id64" id="id65">[18]</a> or an undo/redo feature, to name
just a few. This additional layer also allows PyTables to be relatively
independent of its underlying libraries (and their possible limitations). For
example, PyTables can support <a class="reference external" href="http://www.hdfgroup.org/HDF5">HDF5</a> <a class="footnote-reference" href="#id2" id="id15">[1]</a> data types like <cite>enumerated</cite> or <cite>time</cite> that
are available in the <a class="reference external" href="http://www.hdfgroup.org/HDF5">HDF5</a> <a class="footnote-reference" href="#id2" id="id16">[1]</a> library but not in the <a class="reference external" href="http://www.numpy.org">NumPy</a> <a class="footnote-reference" href="#id20" id="id28">[3]</a> package; or even
perform powerful complex queries that are not implemented directly in neither
<a class="reference external" href="http://www.hdfgroup.org/HDF5">HDF5</a> <a class="footnote-reference" href="#id2" id="id17">[1]</a> nor <a class="reference external" href="http://www.numpy.org">NumPy</a> <a class="footnote-reference" href="#id20" id="id29">[3]</a>.</p>
<p>Furthermore, PyTables also tries hard to be a high performance interface to
HDF5/NumPy, implementing niceties like internal LRU caches for nodes and other
data and metadata, <a class="reference internal" href="usersguide/optimization.html#chunksizefinetune"><em>automatic computation of optimal chunk sizes</em></a> for the datasets, a variety of compressors, ranging from
slow but efficient (<a class="reference external" href="http://www.bzip.org">bzip2</a> <a class="footnote-reference" href="#id66" id="id67">[19]</a>) to extremely fast ones (<a class="reference external" href="http://blosc.pytables.org">Blosc</a> <a class="footnote-reference" href="#id68" id="id69">[20]</a>) in addition to the
standard <a class="reference external" href="http://zlib.net">zlib</a> <a class="footnote-reference" href="#id70" id="id71">[21]</a>. Another difference is that PyTables makes use of <a class="reference external" href="http://code.google.com/p/numexpr">numexpr</a> <a class="footnote-reference" href="#id72" id="id73">[22]</a> so
as to accelerate internal computations (for example, in evaluating complex
queries) to a maximum.</p>
<p>For contrasting with other opinions, you may want to check the PyTables/h5py
comparison in a similar entry of the <a class="reference external" href="http://code.google.com/p/h5py/wiki/FAQ">FAQ of h5py</a> <a class="footnote-reference" href="#id76" id="id77">[23]</a>.</p>
</div>
<div class="section" id="i-ve-found-a-bug-what-do-i-do">
<h3>I’ve found a bug. What do I do?<a class="headerlink" href="#i-ve-found-a-bug-what-do-i-do" title="Permalink to this headline">¶</a></h3>
<p>The PyTables development team works hard to make this eventuality as rare as
possible, but, as in any software made by human beings, bugs do occur. If you
find any bug, please tell us by file a bug report in the <a class="reference external" href="https://github.com/PyTables/PyTables/issues">issue tracker</a> <a class="footnote-reference" href="#id78" id="id79">[24]</a> on
<a class="reference external" href="https://github.com">GitHub</a> <a class="footnote-reference" href="#id80" id="id81">[25]</a>.</p>
</div>
<div class="section" id="is-it-possible-to-get-involved-in-pytables-development">
<h3>Is it possible to get involved in PyTables development?<a class="headerlink" href="#is-it-possible-to-get-involved-in-pytables-development" title="Permalink to this headline">¶</a></h3>
<p>Indeed. We are keen for more people to help out contributing code, unit tests,
documentation, and helping out maintaining this wiki. Drop us a mail on the
<cite>users mailing list</cite> and tell us in which area do you want to work.</p>
</div>
<div class="section" id="how-can-i-cite-pytables">
<h3>How can I cite PyTables?<a class="headerlink" href="#how-can-i-cite-pytables" title="Permalink to this headline">¶</a></h3>
<p>The recommended way to cite PyTables in a paper or a presentation is as
following:</p>
<ul class="simple">
<li>Author: Francesc Alted, Ivan Vilata and others</li>
<li>Title: PyTables: Hierarchical Datasets in Python</li>
<li>Year: 2002 -</li>
<li>URL: <a class="reference external" href="http://www.pytables.org">http://www.pytables.org</a></li>
</ul>
<p>Here’s an example of a BibTeX entry:</p>
<div class="highlight-python"><pre>@Misc{,
author = {Francesc Alted and Ivan Vilata and others},
title = {{PyTables}: Hierarchical Datasets in {Python}},
year = {2002--},
url = "http://www.pytables.org/"
}</pre>
</div>
</div>
</div>
<div class="section" id="pytables-2-x-issues">
<h2>PyTables 2.x issues<a class="headerlink" href="#pytables-2-x-issues" title="Permalink to this headline">¶</a></h2>
<div class="section" id="i-m-having-problems-migrating-my-apps-from-pytables-1-x-into-pytables-2-x-please-help">
<h3>I’m having problems migrating my apps from PyTables 1.x into PyTables 2.x. Please, help!<a class="headerlink" href="#i-m-having-problems-migrating-my-apps-from-pytables-1-x-into-pytables-2-x-please-help" title="Permalink to this headline">¶</a></h3>
<p>Sure. However, you should first check out the <a class="reference internal" href="MIGRATING_TO_2.x.html"><em>Migrating from PyTables 1.x to 2.x</em></a>
document.
It should provide hints to the most frequently asked questions on this regard.</p>
</div>
<div class="section" id="for-combined-searches-like-table-where-x-5-x-3-why-was-a-operator-chosen-instead-of-an-and">
<h3>For combined searches like <cite>table.where(‘(x<5) & (x>3)’)</cite>, why was a <cite>&</cite> operator chosen instead of an <cite>and</cite>?<a class="headerlink" href="#for-combined-searches-like-table-where-x-5-x-3-why-was-a-operator-chosen-instead-of-an-and" title="Permalink to this headline">¶</a></h3>
<p>Search expressions are in fact Python expressions written as strings, and they
are evaluated as such. This has the advantage of not having to learn a new
syntax, but it also implies some limitations with logical <cite>and</cite> and <cite>or</cite>
operators, namely that they can not be overloaded in Python. Thus, it is
impossible right now to get an element-wise operation out of an expression like
<cite>‘array1 and array2’</cite>. That’s why one has to choose some other operator, being
<cite>&</cite> and <cite>|</cite> the most similar to their C counterparts <cite>&&</cite> and <cite>||</cite>, which
aren’t available in Python either.</p>
<p>You should be careful about expressions like <cite>‘x<5 & x>3’</cite> and others like <cite>‘3
< x < 5’</cite> which ‘’won’t work as expected’‘, because of the different operator
precedence and the absence of an overloaded logical <cite>and</cite> operator. More on
this in the appendix about condition syntax in the <a class="reference external" href="http://www.hdfgroup.org/HDF5/doc/RM/RM_H5T.html">HDF5 manual</a> <a class="footnote-reference" href="#id82" id="id83">[26]</a>.</p>
<p>There are quite a few packages affected by those limitations including <a class="reference external" href="http://www.numpy.org">NumPy</a> <a class="footnote-reference" href="#id20" id="id30">[3]</a>
themselves and <a class="reference external" href="http://sqlobject.org">SQLObject</a> <a class="footnote-reference" href="#id84" id="id85">[27]</a>, and there have been quite longish discussions about
adding the possibility of overloading logical operators to Python (see <a class="reference external" href="http://www.python.org/dev/peps/pep-0335">PEP
335</a> <a class="footnote-reference" href="#id86" id="id87">[28]</a> and <a class="reference external" href="http://mail.python.org/pipermail/python-dev/2004-September/048763.html">this thread</a> <a class="footnote-reference" href="#id90" id="id91">[30]</a> for more details).</p>
</div>
<div class="section" id="i-can-not-select-rows-using-in-kernel-queries-with-a-condition-that-involves-an-uint64col-why">
<h3>I can not select rows using in-kernel queries with a condition that involves an UInt64Col. Why?<a class="headerlink" href="#i-can-not-select-rows-using-in-kernel-queries-with-a-condition-that-involves-an-uint64col-why" title="Permalink to this headline">¶</a></h3>
<p>This turns out to be a limitation of the <a class="reference external" href="http://code.google.com/p/numexpr">numexpr</a> <a class="footnote-reference" href="#id72" id="id74">[22]</a> package. Internally,
<a class="reference external" href="http://code.google.com/p/numexpr">numexpr</a> <a class="footnote-reference" href="#id72" id="id75">[22]</a> uses a limited set of types for doing calculations, and unsigned
integers are always upcasted to the immediate signed integer that can fit the
information. The problem here is that there is not a (standard) signed integer
that can be used to keep the information of a 64-bit unsigned integer.</p>
<p>So, your best bet right now is to avoid <cite>uint64</cite> types if you can. If you
absolutely need <cite>uint64</cite>, the only way for doing selections with this is
through regular Python selections. For example, if your table has a <cite>colM</cite>
column which is declared as an <cite>UInt64Col</cite>, then you can still filter its
values with:</p>
<div class="highlight-python"><div class="highlight"><pre><span class="p">[</span><span class="n">row</span><span class="p">[</span><span class="s">'colN'</span><span class="p">]</span> <span class="k">for</span> <span class="n">row</span> <span class="ow">in</span> <span class="n">table</span> <span class="k">if</span> <span class="n">row</span><span class="p">[</span><span class="s">'colM'</span><span class="p">]</span> <span class="o"><</span> <span class="n">X</span><span class="p">]</span>
</pre></div>
</div>
<p>However, this approach will generally lead to slow speed (specially on Win32
platforms, where the values will be converted to Python <cite>long</cite> values).</p>
</div>
<div class="section" id="i-m-already-using-pytables-2-x-but-i-m-still-getting-numarray-objects-instead-of-numpy-ones">
<h3>I’m already using PyTables 2.x but I’m still getting numarray objects instead of NumPy ones!<a class="headerlink" href="#i-m-already-using-pytables-2-x-but-i-m-still-getting-numarray-objects-instead-of-numpy-ones" title="Permalink to this headline">¶</a></h3>
<p>This is most probably due to the fact that you are using a file created with
PyTables 1.x series. By default, PyTables 1.x was setting an HDF5 attribute
<cite>FLAVOR</cite> with the value <cite>‘numarray’</cite> to all leaves. Now, PyTables 2.x sees
this attribute and obediently converts the internal object (truly a NumPy
object) into a <cite>numarray</cite> one. For PyTables 2.x files the <cite>FLAVOR</cite> attribute
will only be saved when explicitly set via the <cite>leaf.flavor</cite> property (or when
passing data to an <tt class="xref py py-class docutils literal"><span class="pre">Array</span></tt> or <tt class="xref py py-class docutils literal"><span class="pre">Table</span></tt> at creation time), so you
will be able to distinguish default flavors from user-set ones by checking the
existence of the <cite>FLAVOR</cite> attribute.</p>
<p>Meanwhile, if you don’t want to receive <cite>numarray</cite> objects when reading old
files, you have several possibilities:</p>
<ul>
<li><p class="first">Remove the flavor for your datasets by hand:</p>
<div class="highlight-python"><div class="highlight"><pre><span class="k">for</span> <span class="n">leaf</span> <span class="ow">in</span> <span class="n">h5file</span><span class="o">.</span><span class="n">walkNodes</span><span class="p">(</span><span class="n">classname</span><span class="o">=</span><span class="s">'Leaf'</span><span class="p">):</span>
<span class="k">del</span> <span class="n">leaf</span><span class="o">.</span><span class="n">flavor</span>
</pre></div>
</div>
</li>
<li><p class="first">Use the :program:’ptrepack` utility with the flag <em class="xref std std-option">--upgrade-flavors</em>
so as to convert all flavors in old files to the default (effectively by
removing the <cite>FLAVOR</cite> attribute).</p>
</li>
<li><p class="first">Remove the <cite>numarray</cite> (and/or <cite>Numeric</cite>) package from your system.
Then PyTables 2.x will return you pure NumPy objects (it can’t be
otherwise!).</p>
</li>
</ul>
</div>
</div>
<div class="section" id="installation-issues">
<h2>Installation issues<a class="headerlink" href="#installation-issues" title="Permalink to this headline">¶</a></h2>
<div class="section" id="windows">
<h3>Windows<a class="headerlink" href="#windows" title="Permalink to this headline">¶</a></h3>
<div class="section" id="error-when-importing-tables">
<h4>Error when importing tables<a class="headerlink" href="#error-when-importing-tables" title="Permalink to this headline">¶</a></h4>
<p>You have installed the binary installer for Windows and, when importing the
<em>tables</em> package you are getting an error like:</p>
<div class="highlight-python"><pre>The command in "0x6714a822" refers to memory in "0x012011a0". The
procedure "written" could not be executed.
Click to ok to terminate.
Click to abort to debug the program.</pre>
</div>
<p>This problem can be due to a series of reasons, but the most probable one is
that you have a version of a DLL library that is needed by PyTables and it is
not at the correct version. Please, double-check the versions of the required
libraries for PyTables and install newer versions, if needed. In most cases,
this solves the issue.</p>
<p>In case you continue getting problems, there are situations where other
programs do install libraries in the PATH that are <strong>optional</strong> to PyTables
(for example BZIP2 or LZO), but that they will be used if they are found in
your system (i.e. anywhere in your <span class="target" id="index-0"></span><tt class="xref std std-envvar docutils literal"><span class="pre">PATH</span></tt>). So, if you find any of
these libraries in your PATH, upgrade it to the latest version available (you
don’t need to re-install PyTables).</p>
</div>
<div class="section" id="can-t-find-lzo-binaries-for-windows">
<h4>Can’t find LZO binaries for Windows<a class="headerlink" href="#can-t-find-lzo-binaries-for-windows" title="Permalink to this headline">¶</a></h4>
<p>Unfortunately, the LZO binaries for Windows seems to be unavailable from its
usual place at <a class="reference external" href="http://gnuwin32.sourceforge.net/packages/lzo.htm">http://gnuwin32.sourceforge.net/packages/lzo.htm</a>. So, in order
to allow people to be able to install this excellent compressor easily, we have
packaged the LZO binaries in a zip file available at:
<a class="reference external" href="http://www.pytables.org/download/lzo-win">http://www.pytables.org/download/lzo-win</a>. This zip file follows the same
structure that a typical <a class="reference external" href="http://gnuwin32.sourceforge.net">GnuWin32</a> <a class="footnote-reference" href="#id88" id="id89">[29]</a> package, so it is just a matter of unpacking
it in your <tt class="docutils literal"><span class="pre">GNUWIN32</span></tt> directory and following the <a class="reference internal" href="usersguide/installation.html#prerequisitesbininst"><em>instructions</em></a> in the <a class="reference external" href="http://www.pytables.org/docs/manual">PyTables Manual</a> <a class="footnote-reference" href="#id58" id="id59">[15]</a>.</p>
<p>Hopefully somebody else will take care again of maintaining LZO for Windows
again.</p>
</div>
</div>
</div>
<div class="section" id="testing-issues">
<h2>Testing issues<a class="headerlink" href="#testing-issues" title="Permalink to this headline">¶</a></h2>
<div class="section" id="tests-fail-when-running-from-ipython">
<h3>Tests fail when running from IPython<a class="headerlink" href="#tests-fail-when-running-from-ipython" title="Permalink to this headline">¶</a></h3>
<p>You may be getting errors related with Doctest when running the test suite from
IPython. This is a known limitation in IPython (see
<a class="reference external" href="http://lists.ipython.scipy.org/pipermail/ipython-dev/2007-April/002859.html">http://lists.ipython.scipy.org/pipermail/ipython-dev/2007-April/002859.html</a>).
Try running the test suite from the vanilla Python interpreter instead.</p>
</div>
<div class="section" id="tests-fail-when-running-from-python-2-5-and-numeric-is-installed">
<h3>Tests fail when running from Python 2.5 and Numeric is installed<a class="headerlink" href="#tests-fail-when-running-from-python-2-5-and-numeric-is-installed" title="Permalink to this headline">¶</a></h3>
<p><cite>Numeric</cite> doesn’t get well with Python 2.5, even on 32-bit platforms. This is
a consequence of <cite>Numeric</cite> not being maintained anymore and you should consider
migrating to NumPy as soon as possible. To get rid of these errors, just
uninstall <cite>Numeric</cite>.</p>
<hr class="docutils" />
<table class="docutils footnote" frame="void" id="id2" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label">[1]</td><td><em>(<a class="fn-backref" href="#id3">1</a>, <a class="fn-backref" href="#id4">2</a>, <a class="fn-backref" href="#id5">3</a>, <a class="fn-backref" href="#id6">4</a>, <a class="fn-backref" href="#id7">5</a>, <a class="fn-backref" href="#id8">6</a>, <a class="fn-backref" href="#id9">7</a>, <a class="fn-backref" href="#id10">8</a>, <a class="fn-backref" href="#id11">9</a>, <a class="fn-backref" href="#id12">10</a>, <a class="fn-backref" href="#id13">11</a>, <a class="fn-backref" href="#id14">12</a>, <a class="fn-backref" href="#id15">13</a>, <a class="fn-backref" href="#id16">14</a>, <a class="fn-backref" href="#id17">15</a>)</em> <a class="reference external" href="http://www.hdfgroup.org/HDF5">http://www.hdfgroup.org/HDF5</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id18" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id19">[2]</a></td><td><a class="reference external" href="http://www.python.org">http://www.python.org</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id20" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label">[3]</td><td><em>(<a class="fn-backref" href="#id21">1</a>, <a class="fn-backref" href="#id22">2</a>, <a class="fn-backref" href="#id23">3</a>, <a class="fn-backref" href="#id24">4</a>, <a class="fn-backref" href="#id25">5</a>, <a class="fn-backref" href="#id26">6</a>, <a class="fn-backref" href="#id27">7</a>, <a class="fn-backref" href="#id28">8</a>, <a class="fn-backref" href="#id29">9</a>, <a class="fn-backref" href="#id30">10</a>)</em> <a class="reference external" href="http://www.numpy.org">http://www.numpy.org</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id31" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id32">[4]</a></td><td><a class="reference external" href="https://lists.sourceforge.net/lists/listinfo/pytables-users">https://lists.sourceforge.net/lists/listinfo/pytables-users</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id33" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id34">[5]</a></td><td><a class="reference external" href="http://sourceforge.net/mailarchive/forum.php?forum_id=13760">http://sourceforge.net/mailarchive/forum.php?forum_id=13760</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id35" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id36">[6]</a></td><td><a class="reference external" href="http://www.mail-archive.com/pytables-users@lists.sourceforge.net/">http://www.mail-archive.com/pytables-users@lists.sourceforge.net/</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id37" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id38">[7]</a></td><td><a class="reference external" href="http://www.hdfgroup.org/HDF5/RD100-2002/">http://www.hdfgroup.org/HDF5/RD100-2002/</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id39" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id40">[8]</a></td><td><a class="reference external" href="http://vitables.org">http://vitables.org</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id41" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label">[9]</td><td><em>(<a class="fn-backref" href="#id42">1</a>, <a class="fn-backref" href="#id43">2</a>)</em> <a class="reference external" href="http://www.cython.org">http://www.cython.org</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id44" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id45">[10]</a></td><td><a class="reference external" href="http://www.pytables.org/moin/FrancescAlted">http://www.pytables.org/moin/FrancescAlted</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id46" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label">[11]</td><td><em>(<a class="fn-backref" href="#id47">1</a>, <a class="fn-backref" href="#id48">2</a>, <a class="fn-backref" href="#id49">3</a>, <a class="fn-backref" href="#id52">4</a>)</em> <a class="reference external" href="http://www.unidata.ucar.edu/software/netcdf">http://www.unidata.ucar.edu/software/netcdf</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id50" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id51">[12]</a></td><td><a class="reference external" href="http://dirac.cnrs-orleans.fr/plone/software/scientificpython">http://dirac.cnrs-orleans.fr/plone/software/scientificpython</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id53" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label">[13]</td><td><em>(<a class="fn-backref" href="#id54">1</a>, <a class="fn-backref" href="#id55">2</a>)</em> <a class="reference external" href="http://www.pydap.org">http://www.pydap.org</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id56" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id57">[14]</a></td><td><a class="reference external" href="http://opendap.org">http://opendap.org</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id58" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id59">[15]</a></td><td><a class="reference external" href="http://www.pytables.org/docs/manual">http://www.pytables.org/docs/manual</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id60" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id61">[16]</a></td><td><a class="reference external" href="http://www.h5py.org">http://www.h5py.org</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id62" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id63">[17]</a></td><td><a class="reference external" href="http://www.pytables.org/moin/ComputingKernel">http://www.pytables.org/moin/ComputingKernel</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id64" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id65">[18]</a></td><td><a class="reference external" href="http://www.pytables.org/moin/PyTablesPro">http://www.pytables.org/moin/PyTablesPro</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id66" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id67">[19]</a></td><td><a class="reference external" href="http://www.bzip.org">http://www.bzip.org</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id68" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id69">[20]</a></td><td><a class="reference external" href="http://blosc.pytables.org">http://blosc.pytables.org</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id70" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id71">[21]</a></td><td><a class="reference external" href="http://zlib.net">http://zlib.net</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id72" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label">[22]</td><td><em>(<a class="fn-backref" href="#id73">1</a>, <a class="fn-backref" href="#id74">2</a>, <a class="fn-backref" href="#id75">3</a>)</em> <a class="reference external" href="http://code.google.com/p/numexpr">http://code.google.com/p/numexpr</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id76" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id77">[23]</a></td><td><a class="reference external" href="http://code.google.com/p/h5py/wiki/FAQ">http://code.google.com/p/h5py/wiki/FAQ</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id78" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id79">[24]</a></td><td><a class="reference external" href="https://github.com/PyTables/PyTables/issues">https://github.com/PyTables/PyTables/issues</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id80" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id81">[25]</a></td><td><a class="reference external" href="https://github.com">https://github.com</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id82" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id83">[26]</a></td><td><a class="reference external" href="http://www.hdfgroup.org/HDF5/doc/RM/RM_H5T.html">http://www.hdfgroup.org/HDF5/doc/RM/RM_H5T.html</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id84" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id85">[27]</a></td><td><a class="reference external" href="http://sqlobject.org">http://sqlobject.org</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id86" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id87">[28]</a></td><td><a class="reference external" href="http://www.python.org/dev/peps/pep-0335">http://www.python.org/dev/peps/pep-0335</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id88" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id89">[29]</a></td><td><a class="reference external" href="http://gnuwin32.sourceforge.net">http://gnuwin32.sourceforge.net</a></td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id90" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id91">[30]</a></td><td><a class="reference external" href="http://mail.python.org/pipermail/python-dev/2004-September/048763.html">http://mail.python.org/pipermail/python-dev/2004-September/048763.html</a></td></tr>
</tbody>
</table>
</div>
</div>
</div>
</div>
</div>
</div>
<div class="sphinxsidebar">
<div class="sphinxsidebarwrapper">
<p class="logo"><a href="index.html">
<img class="logo" src="_static/logo-pytables-small.png" alt="Logo"/>
</a></p>
<h3><a href="index.html">Table Of Contents</a></h3>
<ul>
<li><a class="reference internal" href="#">FAQ</a><ul>
<li><a class="reference internal" href="#general-questions">General questions</a><ul>
<li><a class="reference internal" href="#what-is-pytables">What is PyTables?</a></li>
<li><a class="reference internal" href="#what-are-pytables-licensing-terms">What are PyTables’ licensing terms?</a></li>
<li><a class="reference internal" href="#i-m-having-problems-how-can-i-get-support">I’m having problems. How can I get support?</a></li>
<li><a class="reference internal" href="#why-hdf5">Why HDF5?</a></li>
<li><a class="reference internal" href="#why-python">Why Python?</a></li>
<li><a class="reference internal" href="#why-numpy">Why NumPy?</a></li>
</ul>
</li>
<li><a class="reference internal" href="#where-can-pytables-be-applied">Where can PyTables be applied?</a><ul>
<li><a class="reference internal" href="#is-pytables-safe">Is PyTables safe?</a></li>
<li><a class="reference internal" href="#can-pytables-be-used-in-concurrent-access-scenarios">Can PyTables be used in concurrent access scenarios?</a></li>
<li><a class="reference internal" href="#what-kind-of-containers-does-pytables-implement">What kind of containers does PyTables implement?</a></li>
<li><a class="reference internal" href="#cool-i-d-like-to-see-some-examples-of-use">Cool! I’d like to see some examples of use.</a></li>
<li><a class="reference internal" href="#can-you-show-me-some-screenshots">Can you show me some screenshots?</a></li>
<li><a class="reference internal" href="#is-pytables-a-replacement-for-a-relational-database">Is PyTables a replacement for a relational database?</a></li>
<li><a class="reference internal" href="#how-can-pytables-be-fast-if-it-is-written-in-an-interpreted-language-like-python">How can PyTables be fast if it is written in an interpreted language like Python?</a></li>
<li><a class="reference internal" href="#if-it-is-designed-to-deal-with-very-large-datasets-then-pytables-should-consume-a-lot-of-memory-shouldn-t-it">If it is designed to deal with very large datasets, then PyTables should consume a lot of memory, shouldn’t it?</a></li>
<li><a class="reference internal" href="#why-was-pytables-born">Why was PyTables born?</a></li>
<li><a class="reference internal" href="#does-pytables-have-a-client-server-interface">Does PyTables have a client-server interface?</a></li>
<li><a class="reference internal" href="#how-does-pytables-compare-with-the-h5py-project">How does PyTables compare with the h5py project?</a></li>
<li><a class="reference internal" href="#i-ve-found-a-bug-what-do-i-do">I’ve found a bug. What do I do?</a></li>
<li><a class="reference internal" href="#is-it-possible-to-get-involved-in-pytables-development">Is it possible to get involved in PyTables development?</a></li>
<li><a class="reference internal" href="#how-can-i-cite-pytables">How can I cite PyTables?</a></li>
</ul>
</li>
<li><a class="reference internal" href="#pytables-2-x-issues">PyTables 2.x issues</a><ul>
<li><a class="reference internal" href="#i-m-having-problems-migrating-my-apps-from-pytables-1-x-into-pytables-2-x-please-help">I’m having problems migrating my apps from PyTables 1.x into PyTables 2.x. Please, help!</a></li>
<li><a class="reference internal" href="#for-combined-searches-like-table-where-x-5-x-3-why-was-a-operator-chosen-instead-of-an-and">For combined searches like <cite>table.where(‘(x<5) & (x>3)’)</cite>, why was a <cite>&</cite> operator chosen instead of an <cite>and</cite>?</a></li>
<li><a class="reference internal" href="#i-can-not-select-rows-using-in-kernel-queries-with-a-condition-that-involves-an-uint64col-why">I can not select rows using in-kernel queries with a condition that involves an UInt64Col. Why?</a></li>
<li><a class="reference internal" href="#i-m-already-using-pytables-2-x-but-i-m-still-getting-numarray-objects-instead-of-numpy-ones">I’m already using PyTables 2.x but I’m still getting numarray objects instead of NumPy ones!</a></li>
</ul>
</li>
<li><a class="reference internal" href="#installation-issues">Installation issues</a><ul>
<li><a class="reference internal" href="#windows">Windows</a><ul>
<li><a class="reference internal" href="#error-when-importing-tables">Error when importing tables</a></li>
<li><a class="reference internal" href="#can-t-find-lzo-binaries-for-windows">Can’t find LZO binaries for Windows</a></li>
</ul>
</li>
</ul>
</li>
<li><a class="reference internal" href="#testing-issues">Testing issues</a><ul>
<li><a class="reference internal" href="#tests-fail-when-running-from-ipython">Tests fail when running from IPython</a></li>
<li><a class="reference internal" href="#tests-fail-when-running-from-python-2-5-and-numeric-is-installed">Tests fail when running from Python 2.5 and Numeric is installed</a></li>
</ul>
</li>
</ul>
</li>
</ul>
<h4>Previous topic</h4>
<p class="topless"><a href="cookbook/inmemory_hdf5_files.html"
title="previous chapter">In-memory HDF5 files</a></p>
<h4>Next topic</h4>
<p class="topless"><a href="other_material.html"
title="next chapter">Other Material</a></p>
<h3>This Page</h3>
<ul class="this-page-menu">
<li><a href="_sources/FAQ.txt"
rel="nofollow">Show Source</a></li>
</ul>
<div id="searchbox" style="display: none">
<h3>Quick search</h3>
<form class="search" action="search.html" method="get">
<input type="text" name="q" />
<input type="submit" value="Go" />
<input type="hidden" name="check_keywords" value="yes" />
<input type="hidden" name="area" value="default" />
</form>
<p class="searchtip" style="font-size: 90%">
Enter search terms or a module, class or function name.
</p>
</div>
<script type="text/javascript">$('#searchbox').show(0);</script>
</div>
</div>
<div class="clearer"></div>
</div>
<div class="relbar-bottom">
<div class="related">
<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="np-modindex.html" title="Python Module Index"
>modules</a> </li>
<li class="right" >
<a href="other_material.html" title="Other Material"
>next</a> </li>
<li class="right" >
<a href="cookbook/inmemory_hdf5_files.html" title="In-memory HDF5 files"
>previous</a> </li>
<li><a href="index.html">PyTables 3.0.0 documentation</a> »</li>
</ul>
</div>
</div>
<div class="footer">
© Copyright 2011-2013, PyTables maintainers.
Created using <a href="http://sphinx.pocoo.org/">Sphinx</a> 1.1.3.
</div>
<!-- cloud_sptheme 1.3 -->
</body>
</html>
