| 1 |
<!-- doc/src/sgml/rangetypes.sgml --> |
| 2 |
|
| 3 |
<sect1 id="rangetypes"> |
| 4 |
<title>Range Types</title> |
| 5 |
|
| 6 |
<indexterm> |
| 7 |
<primary>range type</primary> |
| 8 |
</indexterm> |
| 9 |
|
| 10 |
<indexterm> |
| 11 |
<primary>multirange type</primary> |
| 12 |
</indexterm> |
| 13 |
|
| 14 |
<para> |
| 15 |
Range types are data types representing a range of values of some |
| 16 |
element type (called the range's <firstterm>subtype</firstterm>). |
| 17 |
For instance, ranges |
| 18 |
of <type>timestamp</type> might be used to represent the ranges of |
| 19 |
time that a meeting room is reserved. In this case the data type |
| 20 |
is <type>tsrange</type> (short for <quote>timestamp range</quote>), |
| 21 |
and <type>timestamp</type> is the subtype. The subtype must have |
| 22 |
a total order so that it is well-defined whether element values are |
| 23 |
within, before, or after a range of values. |
| 24 |
</para> |
| 25 |
|
| 26 |
<para> |
| 27 |
Range types are useful because they represent many element values in a |
| 28 |
single range value, and because concepts such as overlapping ranges can |
| 29 |
be expressed clearly. The use of time and date ranges for scheduling |
| 30 |
purposes is the clearest example; but price ranges, measurement |
| 31 |
ranges from an instrument, and so forth can also be useful. |
| 32 |
</para> |
| 33 |
|
| 34 |
<para> |
| 35 |
Every range type has a corresponding multirange type. A multirange is |
| 36 |
an ordered list of non-contiguous, non-empty, non-null ranges. Most |
| 37 |
range operators also work on multiranges, and they have a few functions |
| 38 |
of their own. |
| 39 |
</para> |
| 40 |
|
| 41 |
<sect2 id="rangetypes-builtin"> |
| 42 |
<title>Built-in Range and Multirange Types</title> |
| 43 |
|
| 44 |
<para> |
| 45 |
PostgreSQL comes with the following built-in range types: |
| 46 |
<itemizedlist> |
| 47 |
<listitem> |
| 48 |
<para> |
| 49 |
<type>int4range</type> — Range of <type>integer</type>, |
| 50 |
<type>int4multirange</type> — corresponding Multirange |
| 51 |
</para> |
| 52 |
</listitem> |
| 53 |
<listitem> |
| 54 |
<para> |
| 55 |
<type>int8range</type> — Range of <type>bigint</type>, |
| 56 |
<type>int8multirange</type> — corresponding Multirange |
| 57 |
</para> |
| 58 |
</listitem> |
| 59 |
<listitem> |
| 60 |
<para> |
| 61 |
<type>numrange</type> — Range of <type>numeric</type>, |
| 62 |
<type>nummultirange</type> — corresponding Multirange |
| 63 |
</para> |
| 64 |
</listitem> |
| 65 |
<listitem> |
| 66 |
<para> |
| 67 |
<type>tsrange</type> — Range of <type>timestamp without time zone</type>, |
| 68 |
<type>tsmultirange</type> — corresponding Multirange |
| 69 |
</para> |
| 70 |
</listitem> |
| 71 |
<listitem> |
| 72 |
<para> |
| 73 |
<type>tstzrange</type> — Range of <type>timestamp with time zone</type>, |
| 74 |
<type>tstzmultirange</type> — corresponding Multirange |
| 75 |
</para> |
| 76 |
</listitem> |
| 77 |
<listitem> |
| 78 |
<para> |
| 79 |
<type>daterange</type> — Range of <type>date</type>, |
| 80 |
<type>datemultirange</type> — corresponding Multirange |
| 81 |
</para> |
| 82 |
</listitem> |
| 83 |
</itemizedlist> |
| 84 |
In addition, you can define your own range types; |
| 85 |
see <xref linkend="sql-createtype"/> for more information. |
| 86 |
</para> |
| 87 |
</sect2> |
| 88 |
|
| 89 |
<sect2 id="rangetypes-examples"> |
| 90 |
<title>Examples</title> |
| 91 |
|
| 92 |
<para> |
| 93 |
<programlisting> |
| 94 |
CREATE TABLE reservation (room int, during tsrange); |
| 95 |
INSERT INTO reservation VALUES |
| 96 |
(1108, '[2010-01-01 14:30, 2010-01-01 15:30)'); |
| 97 |
|
| 98 |
-- Containment |
| 99 |
SELECT int4range(10, 20) @> 3; |
| 100 |
|
| 101 |
-- Overlaps |
| 102 |
SELECT numrange(11.1, 22.2) && numrange(20.0, 30.0); |
| 103 |
|
| 104 |
-- Extract the upper bound |
| 105 |
SELECT upper(int8range(15, 25)); |
| 106 |
|
| 107 |
-- Compute the intersection |
| 108 |
SELECT int4range(10, 20) * int4range(15, 25); |
| 109 |
|
| 110 |
-- Is the range empty? |
| 111 |
SELECT isempty(numrange(1, 5)); |
| 112 |
</programlisting> |
| 113 |
|
| 114 |
See <xref linkend="range-operators-table"/> |
| 115 |
and <xref linkend="range-functions-table"/> for complete lists of |
| 116 |
operators and functions on range types. |
| 117 |
</para> |
| 118 |
</sect2> |
| 119 |
|
| 120 |
<sect2 id="rangetypes-inclusivity"> |
| 121 |
<title>Inclusive and Exclusive Bounds</title> |
| 122 |
|
| 123 |
<para> |
| 124 |
Every non-empty range has two bounds, the lower bound and the upper |
| 125 |
bound. All points between these values are included in the range. An |
| 126 |
inclusive bound means that the boundary point itself is included in |
| 127 |
the range as well, while an exclusive bound means that the boundary |
| 128 |
point is not included in the range. |
| 129 |
</para> |
| 130 |
|
| 131 |
<para> |
| 132 |
In the text form of a range, an inclusive lower bound is represented by |
| 133 |
<quote><literal>[</literal></quote> while an exclusive lower bound is |
| 134 |
represented by <quote><literal>(</literal></quote>. Likewise, an inclusive upper bound is represented by |
| 135 |
<quote><literal>]</literal></quote>, while an exclusive upper bound is |
| 136 |
represented by <quote><literal>)</literal></quote>. |
| 137 |
(See <xref linkend="rangetypes-io"/> for more details.) |
| 138 |
</para> |
| 139 |
|
| 140 |
<para> |
| 141 |
The functions <literal>lower_inc</literal> |
| 142 |
and <literal>upper_inc</literal> test the inclusivity of the lower |
| 143 |
and upper bounds of a range value, respectively. |
| 144 |
</para> |
| 145 |
</sect2> |
| 146 |
|
| 147 |
<sect2 id="rangetypes-infinite"> |
| 148 |
<title>Infinite (Unbounded) Ranges</title> |
| 149 |
|
| 150 |
<para> |
| 151 |
The lower bound of a range can be omitted, meaning that all |
| 152 |
values less than the upper bound are included in the range, e.g., |
| 153 |
<literal>(,3]</literal>. Likewise, if the upper bound of the range |
| 154 |
is omitted, then all values greater than the lower bound are included |
| 155 |
in the range. If both lower and upper bounds are omitted, all values |
| 156 |
of the element type are considered to be in the range. Specifying a |
| 157 |
missing bound as inclusive is automatically converted to exclusive, |
| 158 |
e.g., <literal>[,]</literal> is converted to <literal>(,)</literal>. |
| 159 |
You can think of these missing values as +/-infinity, but they are |
| 160 |
special range type values and are considered to be beyond any range |
| 161 |
element type's +/-infinity values. |
| 162 |
</para> |
| 163 |
|
| 164 |
<para> |
| 165 |
Element types that have the notion of <quote>infinity</quote> can |
| 166 |
use them as explicit bound values. For example, with timestamp |
| 167 |
ranges, <literal>[today,infinity)</literal> excludes the special |
| 168 |
<type>timestamp</type> value <literal>infinity</literal>, |
| 169 |
while <literal>[today,infinity]</literal> include it, as does |
| 170 |
<literal>[today,)</literal> and <literal>[today,]</literal>. |
| 171 |
</para> |
| 172 |
|
| 173 |
<para> |
| 174 |
The functions <literal>lower_inf</literal> |
| 175 |
and <literal>upper_inf</literal> test for infinite lower |
| 176 |
and upper bounds of a range, respectively. |
| 177 |
</para> |
| 178 |
</sect2> |
| 179 |
|
| 180 |
<sect2 id="rangetypes-io"> |
| 181 |
<title>Range Input/Output</title> |
| 182 |
|
| 183 |
<para> |
| 184 |
The input for a range value must follow one of the following patterns: |
| 185 |
<synopsis> |
| 186 |
(<replaceable>lower-bound</replaceable>,<replaceable>upper-bound</replaceable>) |
| 187 |
(<replaceable>lower-bound</replaceable>,<replaceable>upper-bound</replaceable>] |
| 188 |
[<replaceable>lower-bound</replaceable>,<replaceable>upper-bound</replaceable>) |
| 189 |
[<replaceable>lower-bound</replaceable>,<replaceable>upper-bound</replaceable>] |
| 190 |
empty |
| 191 |
</synopsis> |
| 192 |
The parentheses or brackets indicate whether the lower and upper bounds |
| 193 |
are exclusive or inclusive, as described previously. |
| 194 |
Notice that the final pattern is <literal>empty</literal>, which |
| 195 |
represents an empty range (a range that contains no points). |
| 196 |
</para> |
| 197 |
|
| 198 |
<para> |
| 199 |
The <replaceable>lower-bound</replaceable> may be either a string |
| 200 |
that is valid input for the subtype, or empty to indicate no |
| 201 |
lower bound. Likewise, <replaceable>upper-bound</replaceable> may be |
| 202 |
either a string that is valid input for the subtype, or empty to |
| 203 |
indicate no upper bound. |
| 204 |
</para> |
| 205 |
|
| 206 |
<para> |
| 207 |
Each bound value can be quoted using <literal>"</literal> (double quote) |
| 208 |
characters. This is necessary if the bound value contains parentheses, |
| 209 |
brackets, commas, double quotes, or backslashes, since these characters |
| 210 |
would otherwise be taken as part of the range syntax. To put a double |
| 211 |
quote or backslash in a quoted bound value, precede it with a |
| 212 |
backslash. (Also, a pair of double quotes within a double-quoted bound |
| 213 |
value is taken to represent a double quote character, analogously to the |
| 214 |
rules for single quotes in SQL literal strings.) Alternatively, you can |
| 215 |
avoid quoting and use backslash-escaping to protect all data characters |
| 216 |
that would otherwise be taken as range syntax. Also, to write a bound |
| 217 |
value that is an empty string, write <literal>""</literal>, since writing |
| 218 |
nothing means an infinite bound. |
| 219 |
</para> |
| 220 |
|
| 221 |
<para> |
| 222 |
Whitespace is allowed before and after the range value, but any whitespace |
| 223 |
between the parentheses or brackets is taken as part of the lower or upper |
| 224 |
bound value. (Depending on the element type, it might or might not be |
| 225 |
significant.) |
| 226 |
</para> |
| 227 |
|
| 228 |
<note> |
| 229 |
<para> |
| 230 |
These rules are very similar to those for writing field values in |
| 231 |
composite-type literals. See <xref linkend="rowtypes-io-syntax"/> for |
| 232 |
additional commentary. |
| 233 |
</para> |
| 234 |
</note> |
| 235 |
|
| 236 |
<para> |
| 237 |
Examples: |
| 238 |
<programlisting> |
| 239 |
-- includes 3, does not include 7, and does include all points in between |
| 240 |
SELECT '[3,7)'::int4range; |
| 241 |
|
| 242 |
-- does not include either 3 or 7, but includes all points in between |
| 243 |
SELECT '(3,7)'::int4range; |
| 244 |
|
| 245 |
-- includes only the single point 4 |
| 246 |
SELECT '[4,4]'::int4range; |
| 247 |
|
| 248 |
-- includes no points (and will be normalized to 'empty') |
| 249 |
SELECT '[4,4)'::int4range; |
| 250 |
</programlisting> |
| 251 |
</para> |
| 252 |
|
| 253 |
<para> |
| 254 |
The input for a multirange is curly brackets (<literal>{</literal> and |
| 255 |
<literal>}</literal>) containing zero or more valid ranges, |
| 256 |
separated by commas. Whitespace is permitted around the brackets and |
| 257 |
commas. This is intended to be reminiscent of array syntax, although |
| 258 |
multiranges are much simpler: they have just one dimension and there is |
| 259 |
no need to quote their contents. (The bounds of their ranges may be |
| 260 |
quoted as above however.) |
| 261 |
</para> |
| 262 |
|
| 263 |
<para> |
| 264 |
Examples: |
| 265 |
<programlisting> |
| 266 |
SELECT '{}'::int4multirange; |
| 267 |
SELECT '{[3,7)}'::int4multirange; |
| 268 |
SELECT '{[3,7), [8,9)}'::int4multirange; |
| 269 |
</programlisting> |
| 270 |
</para> |
| 271 |
|
| 272 |
</sect2> |
| 273 |
|
| 274 |
<sect2 id="rangetypes-construct"> |
| 275 |
<title>Constructing Ranges and Multiranges</title> |
| 276 |
|
| 277 |
<para> |
| 278 |
Each range type has a constructor function with the same name as the range |
| 279 |
type. Using the constructor function is frequently more convenient than |
| 280 |
writing a range literal constant, since it avoids the need for extra |
| 281 |
quoting of the bound values. The constructor function |
| 282 |
accepts two or three arguments. The two-argument form constructs a range |
| 283 |
in standard form (lower bound inclusive, upper bound exclusive), while |
| 284 |
the three-argument form constructs a range with bounds of the form |
| 285 |
specified by the third argument. |
| 286 |
The third argument must be one of the strings |
| 287 |
<quote><literal>()</literal></quote>, |
| 288 |
<quote><literal>(]</literal></quote>, |
| 289 |
<quote><literal>[)</literal></quote>, or |
| 290 |
<quote><literal>[]</literal></quote>. |
| 291 |
For example: |
| 292 |
|
| 293 |
<programlisting> |
| 294 |
-- The full form is: lower bound, upper bound, and text argument indicating |
| 295 |
-- inclusivity/exclusivity of bounds. |
| 296 |
SELECT numrange(1.0, 14.0, '(]'); |
| 297 |
|
| 298 |
-- If the third argument is omitted, '[)' is assumed. |
| 299 |
SELECT numrange(1.0, 14.0); |
| 300 |
|
| 301 |
-- Although '(]' is specified here, on display the value will be converted to |
| 302 |
-- canonical form, since int8range is a discrete range type (see below). |
| 303 |
SELECT int8range(1, 14, '(]'); |
| 304 |
|
| 305 |
-- Using NULL for either bound causes the range to be unbounded on that side. |
| 306 |
SELECT numrange(NULL, 2.2); |
| 307 |
</programlisting> |
| 308 |
</para> |
| 309 |
|
| 310 |
<para> |
| 311 |
Each range type also has a multirange constructor with the same name as the |
| 312 |
multirange type. The constructor function takes zero or more arguments |
| 313 |
which are all ranges of the appropriate type. |
| 314 |
For example: |
| 315 |
|
| 316 |
<programlisting> |
| 317 |
SELECT nummultirange(); |
| 318 |
SELECT nummultirange(numrange(1.0, 14.0)); |
| 319 |
SELECT nummultirange(numrange(1.0, 14.0), numrange(20.0, 25.0)); |
| 320 |
</programlisting> |
| 321 |
</para> |
| 322 |
</sect2> |
| 323 |
|
| 324 |
<sect2 id="rangetypes-discrete"> |
| 325 |
<title>Discrete Range Types</title> |
| 326 |
|
| 327 |
<para> |
| 328 |
A discrete range is one whose element type has a well-defined |
| 329 |
<quote>step</quote>, such as <type>integer</type> or <type>date</type>. |
| 330 |
In these types two elements can be said to be adjacent, when there are |
| 331 |
no valid values between them. This contrasts with continuous ranges, |
| 332 |
where it's always (or almost always) possible to identify other element |
| 333 |
values between two given values. For example, a range over the |
| 334 |
<type>numeric</type> type is continuous, as is a range over <type>timestamp</type>. |
| 335 |
(Even though <type>timestamp</type> has limited precision, and so could |
| 336 |
theoretically be treated as discrete, it's better to consider it continuous |
| 337 |
since the step size is normally not of interest.) |
| 338 |
</para> |
| 339 |
|
| 340 |
<para> |
| 341 |
Another way to think about a discrete range type is that there is a clear |
| 342 |
idea of a <quote>next</quote> or <quote>previous</quote> value for each element value. |
| 343 |
Knowing that, it is possible to convert between inclusive and exclusive |
| 344 |
representations of a range's bounds, by choosing the next or previous |
| 345 |
element value instead of the one originally given. |
| 346 |
For example, in an integer range type <literal>[4,8]</literal> and |
| 347 |
<literal>(3,9)</literal> denote the same set of values; but this would not be so |
| 348 |
for a range over numeric. |
| 349 |
</para> |
| 350 |
|
| 351 |
<para> |
| 352 |
A discrete range type should have a <firstterm>canonicalization</firstterm> |
| 353 |
function that is aware of the desired step size for the element type. |
| 354 |
The canonicalization function is charged with converting equivalent values |
| 355 |
of the range type to have identical representations, in particular |
| 356 |
consistently inclusive or exclusive bounds. |
| 357 |
If a canonicalization function is not specified, then ranges with different |
| 358 |
formatting will always be treated as unequal, even though they might |
| 359 |
represent the same set of values in reality. |
| 360 |
</para> |
| 361 |
|
| 362 |
<para> |
| 363 |
The built-in range types <type>int4range</type>, <type>int8range</type>, |
| 364 |
and <type>daterange</type> all use a canonical form that includes |
| 365 |
the lower bound and excludes the upper bound; that is, |
| 366 |
<literal>[)</literal>. User-defined range types can use other conventions, |
| 367 |
however. |
| 368 |
</para> |
| 369 |
</sect2> |
| 370 |
|
| 371 |
<sect2 id="rangetypes-defining"> |
| 372 |
<title>Defining New Range Types</title> |
| 373 |
|
| 374 |
<para> |
| 375 |
Users can define their own range types. The most common reason to do |
| 376 |
this is to use ranges over subtypes not provided among the built-in |
| 377 |
range types. |
| 378 |
For example, to define a new range type of subtype <type>float8</type>: |
| 379 |
|
| 380 |
<programlisting> |
| 381 |
CREATE TYPE floatrange AS RANGE ( |
| 382 |
subtype = float8, |
| 383 |
subtype_diff = float8mi |
| 384 |
); |
| 385 |
|
| 386 |
SELECT '[1.234, 5.678]'::floatrange; |
| 387 |
</programlisting> |
| 388 |
|
| 389 |
Because <type>float8</type> has no meaningful |
| 390 |
<quote>step</quote>, we do not define a canonicalization |
| 391 |
function in this example. |
| 392 |
</para> |
| 393 |
|
| 394 |
<para> |
| 395 |
When you define your own range you automatically get a corresponding |
| 396 |
multirange type. |
| 397 |
</para> |
| 398 |
|
| 399 |
<para> |
| 400 |
Defining your own range type also allows you to specify a different |
| 401 |
subtype B-tree operator class or collation to use, so as to change the sort |
| 402 |
ordering that determines which values fall into a given range. |
| 403 |
</para> |
| 404 |
|
| 405 |
<para> |
| 406 |
If the subtype is considered to have discrete rather than continuous |
| 407 |
values, the <command>CREATE TYPE</command> command should specify a |
| 408 |
<literal>canonical</literal> function. |
| 409 |
The canonicalization function takes an input range value, and must return |
| 410 |
an equivalent range value that may have different bounds and formatting. |
| 411 |
The canonical output for two ranges that represent the same set of values, |
| 412 |
for example the integer ranges <literal>[1, 7]</literal> and <literal>[1, |
| 413 |
8)</literal>, must be identical. It doesn't matter which representation |
| 414 |
you choose to be the canonical one, so long as two equivalent values with |
| 415 |
different formattings are always mapped to the same value with the same |
| 416 |
formatting. In addition to adjusting the inclusive/exclusive bounds |
| 417 |
format, a canonicalization function might round off boundary values, in |
| 418 |
case the desired step size is larger than what the subtype is capable of |
| 419 |
storing. For instance, a range type over <type>timestamp</type> could be |
| 420 |
defined to have a step size of an hour, in which case the canonicalization |
| 421 |
function would need to round off bounds that weren't a multiple of an hour, |
| 422 |
or perhaps throw an error instead. |
| 423 |
</para> |
| 424 |
|
| 425 |
<para> |
| 426 |
In addition, any range type that is meant to be used with GiST or SP-GiST |
| 427 |
indexes should define a subtype difference, or <literal>subtype_diff</literal>, |
| 428 |
function. (The index will still work without <literal>subtype_diff</literal>, |
| 429 |
but it is likely to be considerably less efficient than if a difference |
| 430 |
function is provided.) The subtype difference function takes two input |
| 431 |
values of the subtype, and returns their difference |
| 432 |
(i.e., <replaceable>X</replaceable> minus <replaceable>Y</replaceable>) represented as |
| 433 |
a <type>float8</type> value. In our example above, the |
| 434 |
function <function>float8mi</function> that underlies the regular <type>float8</type> |
| 435 |
minus operator can be used; but for any other subtype, some type |
| 436 |
conversion would be necessary. Some creative thought about how to |
| 437 |
represent differences as numbers might be needed, too. To the greatest |
| 438 |
extent possible, the <literal>subtype_diff</literal> function should agree with |
| 439 |
the sort ordering implied by the selected operator class and collation; |
| 440 |
that is, its result should be positive whenever its first argument is |
| 441 |
greater than its second according to the sort ordering. |
| 442 |
</para> |
| 443 |
|
| 444 |
<para> |
| 445 |
A less-oversimplified example of a <literal>subtype_diff</literal> function is: |
| 446 |
</para> |
| 447 |
|
| 448 |
<programlisting> |
| 449 |
CREATE FUNCTION time_subtype_diff(x time, y time) RETURNS float8 AS |
| 450 |
'SELECT EXTRACT(EPOCH FROM (x - y))' LANGUAGE sql STRICT IMMUTABLE; |
| 451 |
|
| 452 |
CREATE TYPE timerange AS RANGE ( |
| 453 |
subtype = time, |
| 454 |
subtype_diff = time_subtype_diff |
| 455 |
); |
| 456 |
|
| 457 |
SELECT '[11:10, 23:00]'::timerange; |
| 458 |
</programlisting> |
| 459 |
|
| 460 |
<para> |
| 461 |
See <xref linkend="sql-createtype"/> for more information about creating |
| 462 |
range types. |
| 463 |
</para> |
| 464 |
</sect2> |
| 465 |
|
| 466 |
<sect2 id="rangetypes-indexing"> |
| 467 |
<title>Indexing</title> |
| 468 |
|
| 469 |
<indexterm> |
| 470 |
<primary>range type</primary> |
| 471 |
<secondary>indexes on</secondary> |
| 472 |
</indexterm> |
| 473 |
|
| 474 |
<para> |
| 475 |
GiST and SP-GiST indexes can be created for table columns of range types. |
| 476 |
GiST indexes can be also created for table columns of multirange types. |
| 477 |
For instance, to create a GiST index: |
| 478 |
<programlisting> |
| 479 |
CREATE INDEX reservation_idx ON reservation USING GIST (during); |
| 480 |
</programlisting> |
| 481 |
A GiST or SP-GiST index on ranges can accelerate queries involving these |
| 482 |
range operators: |
| 483 |
<literal>=</literal>, |
| 484 |
<literal>&&</literal>, |
| 485 |
<literal><@</literal>, |
| 486 |
<literal>@></literal>, |
| 487 |
<literal><<</literal>, |
| 488 |
<literal>>></literal>, |
| 489 |
<literal>-|-</literal>, |
| 490 |
<literal>&<</literal>, and |
| 491 |
<literal>&></literal>. |
| 492 |
A GiST index on multiranges can accelerate queries involving the same |
| 493 |
set of multirange operators. |
| 494 |
A GiST index on ranges and GiST index on multiranges can also accelerate |
| 495 |
queries involving these cross-type range to multirange and multirange to |
| 496 |
range operators correspondingly: |
| 497 |
<literal>&&</literal>, |
| 498 |
<literal><@</literal>, |
| 499 |
<literal>@></literal>, |
| 500 |
<literal><<</literal>, |
| 501 |
<literal>>></literal>, |
| 502 |
<literal>-|-</literal>, |
| 503 |
<literal>&<</literal>, and |
| 504 |
<literal>&></literal>. |
| 505 |
See <xref linkend="range-operators-table"/> for more information. |
| 506 |
</para> |
| 507 |
|
| 508 |
<para> |
| 509 |
In addition, B-tree and hash indexes can be created for table columns of |
| 510 |
range types. For these index types, basically the only useful range |
| 511 |
operation is equality. There is a B-tree sort ordering defined for range |
| 512 |
values, with corresponding <literal><</literal> and <literal>></literal> operators, |
| 513 |
but the ordering is rather arbitrary and not usually useful in the real |
| 514 |
world. Range types' B-tree and hash support is primarily meant to |
| 515 |
allow sorting and hashing internally in queries, rather than creation of |
| 516 |
actual indexes. |
| 517 |
</para> |
| 518 |
</sect2> |
| 519 |
|
| 520 |
<sect2 id="rangetypes-constraint"> |
| 521 |
<title>Constraints on Ranges</title> |
| 522 |
|
| 523 |
<indexterm> |
| 524 |
<primary>range type</primary> |
| 525 |
<secondary>exclude</secondary> |
| 526 |
</indexterm> |
| 527 |
|
| 528 |
<para> |
| 529 |
While <literal>UNIQUE</literal> is a natural constraint for scalar |
| 530 |
values, it is usually unsuitable for range types. Instead, an |
| 531 |
exclusion constraint is often more appropriate |
| 532 |
(see <link linkend="sql-createtable-exclude">CREATE TABLE |
| 533 |
... CONSTRAINT ... EXCLUDE</link>). Exclusion constraints allow the |
| 534 |
specification of constraints such as <quote>non-overlapping</quote> on a |
| 535 |
range type. For example: |
| 536 |
|
| 537 |
<programlisting> |
| 538 |
CREATE TABLE reservation ( |
| 539 |
during tsrange, |
| 540 |
EXCLUDE USING GIST (during WITH &&) |
| 541 |
); |
| 542 |
</programlisting> |
| 543 |
|
| 544 |
That constraint will prevent any overlapping values from existing |
| 545 |
in the table at the same time: |
| 546 |
|
| 547 |
<programlisting> |
| 548 |
INSERT INTO reservation VALUES |
| 549 |
('[2010-01-01 11:30, 2010-01-01 15:00)'); |
| 550 |
INSERT 0 1 |
| 551 |
|
| 552 |
INSERT INTO reservation VALUES |
| 553 |
('[2010-01-01 14:45, 2010-01-01 15:45)'); |
| 554 |
ERROR: conflicting key value violates exclusion constraint "reservation_during_excl" |
| 555 |
DETAIL: Key (during)=(["2010-01-01 14:45:00","2010-01-01 15:45:00")) conflicts |
| 556 |
with existing key (during)=(["2010-01-01 11:30:00","2010-01-01 15:00:00")). |
| 557 |
</programlisting> |
| 558 |
</para> |
| 559 |
|
| 560 |
<para> |
| 561 |
You can use the <link linkend="btree-gist"><literal>btree_gist</literal></link> |
| 562 |
extension to define exclusion constraints on plain scalar data types, which |
| 563 |
can then be combined with range exclusions for maximum flexibility. For |
| 564 |
example, after <literal>btree_gist</literal> is installed, the following |
| 565 |
constraint will reject overlapping ranges only if the meeting room numbers |
| 566 |
are equal: |
| 567 |
|
| 568 |
<programlisting> |
| 569 |
CREATE EXTENSION btree_gist; |
| 570 |
CREATE TABLE room_reservation ( |
| 571 |
room text, |
| 572 |
during tsrange, |
| 573 |
EXCLUDE USING GIST (room WITH =, during WITH &&) |
| 574 |
); |
| 575 |
|
| 576 |
INSERT INTO room_reservation VALUES |
| 577 |
('123A', '[2010-01-01 14:00, 2010-01-01 15:00)'); |
| 578 |
INSERT 0 1 |
| 579 |
|
| 580 |
INSERT INTO room_reservation VALUES |
| 581 |
('123A', '[2010-01-01 14:30, 2010-01-01 15:30)'); |
| 582 |
ERROR: conflicting key value violates exclusion constraint "room_reservation_room_during_excl" |
| 583 |
DETAIL: Key (room, during)=(123A, ["2010-01-01 14:30:00","2010-01-01 15:30:00")) conflicts |
| 584 |
with existing key (room, during)=(123A, ["2010-01-01 14:00:00","2010-01-01 15:00:00")). |
| 585 |
|
| 586 |
INSERT INTO room_reservation VALUES |
| 587 |
('123B', '[2010-01-01 14:30, 2010-01-01 15:30)'); |
| 588 |
INSERT 0 1 |
| 589 |
</programlisting> |
| 590 |
</para> |
| 591 |
</sect2> |
| 592 |
</sect1> |