blob: c3f5eb49ce078c416b85141a34fe30140452a0a2 [file] [log] [blame]
<?xml version="1.0" encoding="UTF-8"?>
<!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" xml:lang="en" lang="en">
<head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />
<meta name="copyright" content="(C) Copyright 2023" />
<meta name="DC.rights.owner" content="(C) Copyright 2023" />
<meta name="DC.Type" content="concept" />
<meta name="DC.Title" content="PARQUET_READ_PAGE_INDEX Query Option" />
<meta name="DC.Relation" scheme="URI" content="../topics/impala_query_options.html" />
<meta name="prodname" content="Impala" />
<meta name="version" content="Impala 3.4.x" />
<meta name="DC.Format" content="XHTML" />
<meta name="DC.Identifier" content="parquet_read_page_index" />
<link rel="stylesheet" type="text/css" href="../commonltr.css" />
<title>PARQUET_READ_PAGE_INDEX Query Option</title>
</head>
<body id="parquet_read_page_index">
<h1 class="title topictitle1" id="ariaid-title1">PARQUET_READ_PAGE_INDEX Query Option</h1>
<div class="body conbody">
<p class="p">
Use the <code class="ph codeph">PARQUET_READ_PAGE_INDEX</code> query option to disable or enable using
the Parquet page index during scans. The page index contains min/max statistics at the
page-level granularity. It can be used to skip pages and rows that do not match the
conditions in the <code class="ph codeph">WHERE</code> clause.
</p>
<p class="p">
This option enables the same optimization as the <code class="ph codeph">PARQUET_READ_STATISTICS</code>
at the finer grained page level.
</p>
<p class="p">
Impala supports filtering based on Parquet statistics:
</p>
<ul class="ul">
<li class="li">
Of the types: Boolean, Integer, Decimal, String, Timestamp
</li>
<li class="li">
For simple predicates of the forms: <code class="ph codeph">&lt;slot&gt; &lt;op&gt; &lt;constant&gt;</code> or
<code class="ph codeph">&lt;constant&gt; &lt;op&gt; &lt;slot&gt;</code>, where <code class="ph codeph">&lt;op&gt;</code> is LT,
LE, GE, GT, and EQ
</li>
</ul>
<div class="p">
The supported values for the query option are:
<ul class="ul">
<li class="li">
<code class="ph codeph">true</code> (<code class="ph codeph">1</code>): Read the page-level statistics from the
Parquet page index during query processing and filter out pages based on the
statistics.
</li>
<li class="li">
<code class="ph codeph">false</code> (<code class="ph codeph">0</code>): Do not use the Parquet page index.
</li>
<li class="li">
Any other values are treated as <code class="ph codeph">false</code>.
</li>
</ul>
</div>
<p class="p">
<strong class="ph b">Type:</strong> Boolean
</p>
<p class="p">
<strong class="ph b">Default:</strong> <code class="ph codeph">TRUE</code>
</p>
</div>
<div class="related-links">
<div class="familylinks">
<div class="parentlink"><strong>Parent topic:</strong> <a class="link" href="../topics/impala_query_options.html">Query Options for the SET Statement</a></div>
</div>
</div></body>
</html>