| <?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 2025" /> |
| <meta name="DC.rights.owner" content="(C) Copyright 2025" /> |
| <meta name="DC.Type" content="concept" /> |
| <meta name="DC.Title" content="FLOAT Data Type" /> |
| <meta name="DC.Relation" scheme="URI" content="../topics/impala_datatypes.html" /> |
| <meta name="prodname" content="Impala" /> |
| <meta name="prodname" content="Impala" /> |
| <meta name="version" content="Impala 3.4.x" /> |
| <meta name="version" content="Impala 3.4.x" /> |
| <meta name="DC.Format" content="XHTML" /> |
| <meta name="DC.Identifier" content="float" /> |
| <link rel="stylesheet" type="text/css" href="../commonltr.css" /> |
| <title>FLOAT Data Type</title> |
| </head> |
| <body id="float"> |
| |
| |
| <h1 class="title topictitle1" id="ariaid-title1">FLOAT Data Type</h1> |
| |
| |
| |
| |
| <div class="body conbody"> |
| |
| <p class="p"> |
| A single precision floating-point data type used in <code class="ph codeph">CREATE TABLE</code> and <code class="ph codeph">ALTER |
| TABLE</code> statements. |
| </p> |
| |
| |
| <p class="p"> |
| <strong class="ph b">Syntax:</strong> |
| </p> |
| |
| |
| <p class="p"> |
| In the column definition of a <code class="ph codeph">CREATE TABLE</code> statement: |
| </p> |
| |
| |
| <pre class="pre codeblock"><code><var class="keyword varname">column_name</var> FLOAT</code></pre> |
| |
| <p class="p"> |
| <strong class="ph b">Range:</strong> 1.40129846432481707e-45 .. 3.40282346638528860e+38, positive or negative |
| </p> |
| |
| |
| <p class="p"> |
| <strong class="ph b">Precision:</strong> 6 to 9 significant digits, depending on usage. The number of significant digits does |
| not depend on the position of the decimal point. |
| </p> |
| |
| |
| <p class="p"> |
| <strong class="ph b">Representation:</strong> The values are stored in 4 bytes, using |
| <a class="xref" href="https://en.wikipedia.org/wiki/Single-precision_floating-point_format" target="_blank">IEEE 754 Single Precision Binary Floating Point</a> format. |
| </p> |
| |
| |
| <p class="p"> |
| <strong class="ph b">Conversions:</strong> Impala automatically converts <code class="ph codeph">FLOAT</code> to more precise |
| <code class="ph codeph">DOUBLE</code> values, but not the other way around. You can use <code class="ph codeph">CAST()</code> to convert |
| <code class="ph codeph">FLOAT</code> values to <code class="ph codeph">TINYINT</code>, <code class="ph codeph">SMALLINT</code>, <code class="ph codeph">INT</code>, |
| <code class="ph codeph">BIGINT</code>, <code class="ph codeph">STRING</code>, <code class="ph codeph">TIMESTAMP</code>, or <code class="ph codeph">BOOLEAN</code>. |
| You can use exponential notation in <code class="ph codeph">FLOAT</code> literals or when casting from |
| <code class="ph codeph">STRING</code>, for example <code class="ph codeph">1.0e6</code> to represent one million. |
| <span class="ph"> Casting an integer or floating-point value |
| <code class="ph codeph">N</code> to <code class="ph codeph">TIMESTAMP</code> produces a value that is |
| <code class="ph codeph">N</code> seconds past the start of the epoch date (January 1, 1970). By |
| default, the result value represents a date and time in the UTC time zone. If the |
| setting <code class="ph codeph">‑‑use_local_tz_for_unix_timestamp_conversions=true</code> |
| is in effect, the resulting <code class="ph codeph">TIMESTAMP</code> represents a date and time in the |
| local time zone. </span> |
| </p> |
| |
| |
| <p class="p"> |
| <strong class="ph b">Usage notes:</strong> |
| </p> |
| |
| |
| |
| <p class="p"> |
| Impala does not evaluate NaN (not a number) as equal to any other numeric values, |
| including other NaN values. For example, the following statement, which evaluates |
| equality between two NaN values, returns <code class="ph codeph">false</code>: |
| </p> |
| |
| |
| <pre class="pre codeblock"><code> |
| SELECT CAST('nan' AS FLOAT)=CAST('nan' AS FLOAT); |
| </code></pre> |
| |
| <p class="p"> |
| <strong class="ph b">Examples:</strong> |
| </p> |
| |
| |
| <pre class="pre codeblock"><code>CREATE TABLE t1 (x FLOAT); |
| SELECT CAST(1000.5 AS FLOAT); |
| </code></pre> |
| |
| <p class="p"> |
| <strong class="ph b">Partitioning:</strong> Because fractional values of this type are not always represented |
| precisely, when this type is used for a partition key column, the underlying HDFS |
| directories might not be named exactly as you expect. Prefer to partition on a |
| <code class="ph codeph">DECIMAL</code> column instead. |
| </p> |
| |
| |
| <p class="p"> |
| <strong class="ph b">HBase considerations:</strong> This data type is fully compatible with HBase tables. |
| </p> |
| |
| |
| <p class="p"> |
| <strong class="ph b">Parquet considerations:</strong> This type is fully compatible with Parquet tables. |
| </p> |
| |
| |
| <p class="p"> |
| <strong class="ph b">Text table considerations:</strong> Values of this type are potentially larger in text |
| tables than in tables using Parquet or other binary formats. |
| </p> |
| |
| |
| |
| |
| <p class="p"> |
| <strong class="ph b">Internal details:</strong> Represented in memory as a 4-byte value. |
| </p> |
| |
| |
| |
| |
| <p class="p"> |
| <strong class="ph b">Column statistics considerations:</strong> Because this type has a fixed size, the maximum |
| and average size fields are always filled in for column statistics, even before you run |
| the <code class="ph codeph">COMPUTE STATS</code> statement. |
| </p> |
| |
| |
| <p class="p"> |
| <strong class="ph b">Restrictions:</strong> |
| </p> |
| |
| |
| |
| |
| <p class="p"> |
| Due to the way arithmetic on <code class="ph codeph">FLOAT</code> and <code class="ph codeph">DOUBLE</code> columns |
| uses high-performance hardware instructions, and distributed queries can perform these |
| operations in different order for each query, results can vary slightly for aggregate |
| function calls such as <code class="ph codeph">SUM()</code> and <code class="ph codeph">AVG()</code> for |
| <code class="ph codeph">FLOAT</code> and <code class="ph codeph">DOUBLE</code> columns, particularly on large data |
| sets where millions or billions of values are summed or averaged. For perfect |
| consistency and repeatability, use the <code class="ph codeph">DECIMAL</code> data type for such |
| operations instead of <code class="ph codeph">FLOAT</code> or <code class="ph codeph">DOUBLE</code>. |
| </p> |
| |
| |
| <p class="p"> |
| The inability to exactly represent certain floating-point values means that |
| <code class="ph codeph">DECIMAL</code> is sometimes a better choice than <code class="ph codeph">DOUBLE</code> or |
| <code class="ph codeph">FLOAT</code> when precision is critical, particularly when transferring data |
| from other database systems that use different representations or file formats. |
| </p> |
| |
| |
| <p class="p"> |
| <strong class="ph b">Kudu considerations:</strong> |
| </p> |
| |
| <p class="p"> |
| Currently, the data types <code class="ph codeph">BOOLEAN</code>, <code class="ph codeph">FLOAT</code>, and |
| <code class="ph codeph">DOUBLE</code> cannot be used for primary key columns in Kudu tables. |
| </p> |
| |
| |
| <p class="p"> |
| <strong class="ph b">Related information:</strong> |
| </p> |
| |
| |
| <p class="p"> |
| <a class="xref" href="impala_literals.html#numeric_literals">Numeric Literals</a>, <a class="xref" href="impala_math_functions.html#math_functions">Impala Mathematical Functions</a>, |
| <a class="xref" href="impala_double.html#double">DOUBLE Data Type</a> |
| </p> |
| |
| </div> |
| |
| <div class="related-links"> |
| <div class="familylinks"> |
| <div class="parentlink"><strong>Parent topic:</strong> <a class="link" href="../topics/impala_datatypes.html">Data Types</a></div> |
| </div> |
| </div></body> |
| </html> |
