1 <?xml version="1.0" encoding="UTF-8" standalone="no"?>
2 <!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>F.39. seg — a datatype for line segments or floating point intervals</title><link rel="stylesheet" type="text/css" href="stylesheet.css" /><link rev="made" href="pgsql-docs@lists.postgresql.org" /><meta name="generator" content="DocBook XSL Stylesheets Vsnapshot" /><link rel="prev" href="postgres-fdw.html" title="F.38. postgres_fdw — access data stored in external PostgreSQL servers" /><link rel="next" href="sepgsql.html" title="F.40. sepgsql — SELinux-, label-based mandatory access control (MAC) security module" /></head><body id="docContent" class="container-fluid col-10"><div class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">F.39. seg — a datatype for line segments or floating point intervals</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="postgres-fdw.html" title="F.38. postgres_fdw — access data stored in external PostgreSQL servers">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="contrib.html" title="Appendix F. Additional Supplied Modules and Extensions">Up</a></td><th width="60%" align="center">Appendix F. Additional Supplied Modules and Extensions</th><td width="10%" align="right"><a accesskey="h" href="index.html" title="PostgreSQL 18.0 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="sepgsql.html" title="F.40. sepgsql — SELinux-, label-based mandatory access control (MAC) security module">Next</a></td></tr></table><hr /></div><div class="sect1" id="SEG"><div class="titlepage"><div><div><h2 class="title" style="clear: both">F.39. seg — a datatype for line segments or floating point intervals <a href="#SEG" class="id_link">#</a></h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="seg.html#SEG-RATIONALE">F.39.1. Rationale</a></span></dt><dt><span class="sect2"><a href="seg.html#SEG-SYNTAX">F.39.2. Syntax</a></span></dt><dt><span class="sect2"><a href="seg.html#SEG-PRECISION">F.39.3. Precision</a></span></dt><dt><span class="sect2"><a href="seg.html#SEG-USAGE">F.39.4. Usage</a></span></dt><dt><span class="sect2"><a href="seg.html#SEG-NOTES">F.39.5. Notes</a></span></dt><dt><span class="sect2"><a href="seg.html#SEG-CREDITS">F.39.6. Credits</a></span></dt></dl></div><a id="id-1.11.7.49.2" class="indexterm"></a><p>
3 This module implements a data type <code class="type">seg</code> for
4 representing line segments, or floating point intervals.
5 <code class="type">seg</code> can represent uncertainty in the interval endpoints,
6 making it especially useful for representing laboratory measurements.
8 This module is considered <span class="quote">“<span class="quote">trusted</span>”</span>, that is, it can be
9 installed by non-superusers who have <code class="literal">CREATE</code> privilege
10 on the current database.
11 </p><div class="sect2" id="SEG-RATIONALE"><div class="titlepage"><div><div><h3 class="title">F.39.1. Rationale <a href="#SEG-RATIONALE" class="id_link">#</a></h3></div></div></div><p>
12 The geometry of measurements is usually more complex than that of a
13 point in a numeric continuum. A measurement is usually a segment of
14 that continuum with somewhat fuzzy limits. The measurements come out
15 as intervals because of uncertainty and randomness, as well as because
16 the value being measured may naturally be an interval indicating some
17 condition, such as the temperature range of stability of a protein.
19 Using just common sense, it appears more convenient to store such data
20 as intervals, rather than pairs of numbers. In practice, it even turns
21 out more efficient in most applications.
23 Further along the line of common sense, the fuzziness of the limits
24 suggests that the use of traditional numeric data types leads to a
25 certain loss of information. Consider this: your instrument reads
26 6.50, and you input this reading into the database. What do you get
27 when you fetch it? Watch:
29 </p><pre class="screen">
30 test=> select 6.50 :: float8 as "pH";
37 In the world of measurements, 6.50 is not the same as 6.5. It may
38 sometimes be critically different. The experimenters usually write
39 down (and publish) the digits they trust. 6.50 is actually a fuzzy
40 interval contained within a bigger and even fuzzier interval, 6.5,
41 with their center points being (probably) the only common feature they
42 share. We definitely do not want such different data items to appear the
45 Conclusion? It is nice to have a special data type that can record the
46 limits of an interval with arbitrarily variable precision. Variable in
47 the sense that each data element records its own precision.
51 </p><pre class="screen">
52 test=> select '6.25 .. 6.50'::seg as "pH";
58 </p></div><div class="sect2" id="SEG-SYNTAX"><div class="titlepage"><div><div><h3 class="title">F.39.2. Syntax <a href="#SEG-SYNTAX" class="id_link">#</a></h3></div></div></div><p>
59 The external representation of an interval is formed using one or two
60 floating-point numbers joined by the range operator (<code class="literal">..</code>
61 or <code class="literal">...</code>). Alternatively, it can be specified as a
62 center point plus or minus a deviation.
63 Optional certainty indicators (<code class="literal"><</code>,
64 <code class="literal">></code> or <code class="literal">~</code>) can be stored as well.
65 (Certainty indicators are ignored by all the built-in operators, however.)
66 <a class="xref" href="seg.html#SEG-REPR-TABLE" title="Table F.29. seg External Representations">Table F.29</a> gives an overview of allowed
67 representations; <a class="xref" href="seg.html#SEG-INPUT-EXAMPLES" title="Table F.30. Examples of Valid seg Input">Table F.30</a> shows some
70 In <a class="xref" href="seg.html#SEG-REPR-TABLE" title="Table F.29. seg External Representations">Table F.29</a>, <em class="replaceable"><code>x</code></em>, <em class="replaceable"><code>y</code></em>, and
71 <em class="replaceable"><code>delta</code></em> denote
72 floating-point numbers. <em class="replaceable"><code>x</code></em> and <em class="replaceable"><code>y</code></em>, but
73 not <em class="replaceable"><code>delta</code></em>, can be preceded by a certainty indicator.
74 </p><div class="table" id="SEG-REPR-TABLE"><p class="title"><strong>Table F.29. <code class="type">seg</code> External Representations</strong></p><div class="table-contents"><table class="table" summary="seg External Representations" border="1"><colgroup><col /><col /></colgroup><tbody><tr><td><code class="literal"><em class="replaceable"><code>x</code></em></code></td><td>Single value (zero-length interval)
75 </td></tr><tr><td><code class="literal"><em class="replaceable"><code>x</code></em> .. <em class="replaceable"><code>y</code></em></code></td><td>Interval from <em class="replaceable"><code>x</code></em> to <em class="replaceable"><code>y</code></em>
76 </td></tr><tr><td><code class="literal"><em class="replaceable"><code>x</code></em> (+-) <em class="replaceable"><code>delta</code></em></code></td><td>Interval from <em class="replaceable"><code>x</code></em> - <em class="replaceable"><code>delta</code></em> to
77 <em class="replaceable"><code>x</code></em> + <em class="replaceable"><code>delta</code></em>
78 </td></tr><tr><td><code class="literal"><em class="replaceable"><code>x</code></em> ..</code></td><td>Open interval with lower bound <em class="replaceable"><code>x</code></em>
79 </td></tr><tr><td><code class="literal">.. <em class="replaceable"><code>x</code></em></code></td><td>Open interval with upper bound <em class="replaceable"><code>x</code></em>
80 </td></tr></tbody></table></div></div><br class="table-break" /><div class="table" id="SEG-INPUT-EXAMPLES"><p class="title"><strong>Table F.30. Examples of Valid <code class="type">seg</code> Input</strong></p><div class="table-contents"><table class="table" summary="Examples of Valid seg Input" border="1"><colgroup><col class="col1" /><col class="col2" /></colgroup><tbody><tr><td><code class="literal">5.0</code></td><td>
81 Creates a zero-length segment (a point, if you will)
82 </td></tr><tr><td><code class="literal">~5.0</code></td><td>
83 Creates a zero-length segment and records
84 <code class="literal">~</code> in the data. <code class="literal">~</code> is ignored
85 by <code class="type">seg</code> operations, but
86 is preserved as a comment.
87 </td></tr><tr><td><code class="literal"><5.0</code></td><td>
88 Creates a point at 5.0. <code class="literal"><</code> is ignored but
89 is preserved as a comment.
90 </td></tr><tr><td><code class="literal">>5.0</code></td><td>
91 Creates a point at 5.0. <code class="literal">></code> is ignored but
92 is preserved as a comment.
93 </td></tr><tr><td><code class="literal">5(+-)0.3</code></td><td>
94 Creates an interval <code class="literal">4.7 .. 5.3</code>.
95 Note that the <code class="literal">(+-)</code> notation isn't preserved.
96 </td></tr><tr><td><code class="literal">50 .. </code></td><td>Everything that is greater than or equal to 50</td></tr><tr><td><code class="literal">.. 0</code></td><td>Everything that is less than or equal to 0</td></tr><tr><td><code class="literal">1.5e-2 .. 2E-2 </code></td><td>Creates an interval <code class="literal">0.015 .. 0.02</code></td></tr><tr><td><code class="literal">1 ... 2</code></td><td>
97 The same as <code class="literal">1...2</code>, or <code class="literal">1 .. 2</code>,
98 or <code class="literal">1..2</code>
99 (spaces around the range operator are ignored)
100 </td></tr></tbody></table></div></div><br class="table-break" /><p>
101 Because the <code class="literal">...</code> operator is widely used in data sources, it is allowed
102 as an alternative spelling of the <code class="literal">..</code> operator. Unfortunately, this
103 creates a parsing ambiguity: it is not clear whether the upper bound
104 in <code class="literal">0...23</code> is meant to be <code class="literal">23</code> or <code class="literal">0.23</code>.
105 This is resolved by requiring at least one digit before the decimal
106 point in all numbers in <code class="type">seg</code> input.
108 As a sanity check, <code class="type">seg</code> rejects intervals with the lower bound
109 greater than the upper, for example <code class="literal">5 .. 2</code>.
110 </p></div><div class="sect2" id="SEG-PRECISION"><div class="titlepage"><div><div><h3 class="title">F.39.3. Precision <a href="#SEG-PRECISION" class="id_link">#</a></h3></div></div></div><p>
111 <code class="type">seg</code> values are stored internally as pairs of 32-bit floating point
112 numbers. This means that numbers with more than 7 significant digits
115 Numbers with 7 or fewer significant digits retain their
116 original precision. That is, if your query returns 0.00, you will be
117 sure that the trailing zeroes are not the artifacts of formatting: they
118 reflect the precision of the original data. The number of leading
119 zeroes does not affect precision: the value 0.0067 is considered to
120 have just 2 significant digits.
121 </p></div><div class="sect2" id="SEG-USAGE"><div class="titlepage"><div><div><h3 class="title">F.39.4. Usage <a href="#SEG-USAGE" class="id_link">#</a></h3></div></div></div><p>
122 The <code class="filename">seg</code> module includes a GiST index operator class for
123 <code class="type">seg</code> values.
124 The operators supported by the GiST operator class are shown in <a class="xref" href="seg.html#SEG-GIST-OPERATORS" title="Table F.31. Seg GiST Operators">Table F.31</a>.
125 </p><div class="table" id="SEG-GIST-OPERATORS"><p class="title"><strong>Table F.31. Seg GiST Operators</strong></p><div class="table-contents"><table class="table" summary="Seg GiST Operators" border="1"><colgroup><col /></colgroup><thead><tr><th class="func_table_entry"><p class="func_signature">
130 </p></th></tr></thead><tbody><tr><td class="func_table_entry"><p class="func_signature">
131 <code class="type">seg</code> <code class="literal"><<</code> <code class="type">seg</code>
132 → <code class="returnvalue">boolean</code>
135 Is the first <code class="type">seg</code> entirely to the left of the second?
136 [a, b] << [c, d] is true if b < c.
137 </p></td></tr><tr><td class="func_table_entry"><p class="func_signature">
138 <code class="type">seg</code> <code class="literal">>></code> <code class="type">seg</code>
139 → <code class="returnvalue">boolean</code>
142 Is the first <code class="type">seg</code> entirely to the right of the second?
143 [a, b] >> [c, d] is true if a > d.
144 </p></td></tr><tr><td class="func_table_entry"><p class="func_signature">
145 <code class="type">seg</code> <code class="literal">&<</code> <code class="type">seg</code>
146 → <code class="returnvalue">boolean</code>
149 Does the first <code class="type">seg</code> not extend to the right of the
151 [a, b] &< [c, d] is true if b <= d.
152 </p></td></tr><tr><td class="func_table_entry"><p class="func_signature">
153 <code class="type">seg</code> <code class="literal">&></code> <code class="type">seg</code>
154 → <code class="returnvalue">boolean</code>
157 Does the first <code class="type">seg</code> not extend to the left of the
159 [a, b] &> [c, d] is true if a >= c.
160 </p></td></tr><tr><td class="func_table_entry"><p class="func_signature">
161 <code class="type">seg</code> <code class="literal">=</code> <code class="type">seg</code>
162 → <code class="returnvalue">boolean</code>
165 Are the two <code class="type">seg</code>s equal?
166 </p></td></tr><tr><td class="func_table_entry"><p class="func_signature">
167 <code class="type">seg</code> <code class="literal">&&</code> <code class="type">seg</code>
168 → <code class="returnvalue">boolean</code>
171 Do the two <code class="type">seg</code>s overlap?
172 </p></td></tr><tr><td class="func_table_entry"><p class="func_signature">
173 <code class="type">seg</code> <code class="literal">@></code> <code class="type">seg</code>
174 → <code class="returnvalue">boolean</code>
177 Does the first <code class="type">seg</code> contain the second?
178 </p></td></tr><tr><td class="func_table_entry"><p class="func_signature">
179 <code class="type">seg</code> <code class="literal"><@</code> <code class="type">seg</code>
180 → <code class="returnvalue">boolean</code>
183 Is the first <code class="type">seg</code> contained in the second?
184 </p></td></tr></tbody></table></div></div><br class="table-break" /><p>
185 In addition to the above operators, the usual comparison
186 operators shown in <a class="xref" href="functions-comparison.html#FUNCTIONS-COMPARISON-OP-TABLE" title="Table 9.1. Comparison Operators">Table 9.1</a> are
187 available for type <code class="type">seg</code>. These operators
188 first compare (a) to (c),
189 and if these are equal, compare (b) to (d). That results in
190 reasonably good sorting in most cases, which is useful if
191 you want to use ORDER BY with this type.
192 </p></div><div class="sect2" id="SEG-NOTES"><div class="titlepage"><div><div><h3 class="title">F.39.5. Notes <a href="#SEG-NOTES" class="id_link">#</a></h3></div></div></div><p>
193 For examples of usage, see the regression test <code class="filename">sql/seg.sql</code>.
195 The mechanism that converts <code class="literal">(+-)</code> to regular ranges
196 isn't completely accurate in determining the number of significant digits
197 for the boundaries. For example, it adds an extra digit to the lower
198 boundary if the resulting interval includes a power of ten:
200 </p><pre class="screen">
201 postgres=> select '10(+-)1'::seg as seg;
204 9.0 .. 11 -- should be: 9 .. 11
207 The performance of an R-tree index can largely depend on the initial
208 order of input values. It may be very helpful to sort the input table
209 on the <code class="type">seg</code> column; see the script <code class="filename">sort-segments.pl</code>
211 </p></div><div class="sect2" id="SEG-CREDITS"><div class="titlepage"><div><div><h3 class="title">F.39.6. Credits <a href="#SEG-CREDITS" class="id_link">#</a></h3></div></div></div><p>
212 Original author: Gene Selkov, Jr. <code class="email"><<a class="email" href="mailto:selkovjr@mcs.anl.gov">selkovjr@mcs.anl.gov</a>></code>,
213 Mathematics and Computer Science Division, Argonne National Laboratory.
215 My thanks are primarily to Prof. Joe Hellerstein
216 (<a class="ulink" href="https://dsf.berkeley.edu/jmh/" target="_top">https://dsf.berkeley.edu/jmh/</a>) for elucidating the
217 gist of the GiST (<a class="ulink" href="http://gist.cs.berkeley.edu/" target="_top">http://gist.cs.berkeley.edu/</a>). I am
218 also grateful to all Postgres developers, present and past, for enabling
219 myself to create my own world and live undisturbed in it. And I would like
220 to acknowledge my gratitude to Argonne Lab and to the U.S. Department of
221 Energy for the years of faithful support of my database research.
222 </p></div></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="postgres-fdw.html" title="F.38. postgres_fdw — access data stored in external PostgreSQL servers">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="contrib.html" title="Appendix F. Additional Supplied Modules and Extensions">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="sepgsql.html" title="F.40. sepgsql — SELinux-, label-based mandatory access control (MAC) security module">Next</a></td></tr><tr><td width="40%" align="left" valign="top">F.38. postgres_fdw —
223 access data stored in external <span class="productname">PostgreSQL</span>
224 servers </td><td width="20%" align="center"><a accesskey="h" href="index.html" title="PostgreSQL 18.0 Documentation">Home</a></td><td width="40%" align="right" valign="top"> F.40. sepgsql —
225 SELinux-, label-based mandatory access control (MAC) security module</td></tr></table></div></body></html>