rdelaage
/
postgresql
mirror of https://git.postgresql.org/git/postgresql.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
postgresql/contrib/cube/sql/cube.sql

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

433 lines
19 KiB
MySQL
Raw Normal View History

Gene Selkov's CUBE datatype (GiST example code)
2000-12-11 21:39:15 +01:00
--
-- Test cube datatype
--
Convert contrib modules to use the extension facility. This isn't fully tested as yet, in particular I'm not sure that the "foo--unpackaged--1.0.sql" scripts are OK. But it's time to get some buildfarm cycles on it. sepgsql is not converted to an extension, mainly because it seems to require a very nonstandard installation process. Dimitri Fontaine and Tom Lane
2011-02-14 02:06:41 +01:00
CREATE EXTENSION cube;
Gene Selkov's CUBE datatype (GiST example code)
2000-12-11 21:39:15 +01:00
Test all contrib-created operator classes with amvalidate. I'd supposed that people would do this manually when creating new operator classes, but the folly of that was exposed today. The tests seem fast enough that we can just apply them during the normal regression tests. contrib/isn fails the checks for lack of complete sets of cross-type operators. That's a nice-to-have policy rather than a functional requirement, so leave it as-is, but insert ORDER BY in the query to ensure consistent cross-platform output. Discussion: https://postgr.es/m/7076.1480446837@sss.pgh.pa.us
2016-11-29 21:05:22 +01:00
-- Check whether any of our opclasses fail amvalidate
SELECT amname, opcname
FROM pg_opclass opc LEFT JOIN pg_am am ON am.oid = opcmethod
WHERE opc.oid >= 16384 AND NOT amvalidate(opc.oid);
Gene Selkov's CUBE datatype (GiST example code)
2000-12-11 21:39:15 +01:00
--
-- testing the input and output functions
--
-- Any number (a one-dimensional point)
SELECT '1'::cube AS cube;
SELECT '-1'::cube AS cube;
SELECT '1.'::cube AS cube;
SELECT '-1.'::cube AS cube;
SELECT '.1'::cube AS cube;
SELECT '-.1'::cube AS cube;
SELECT '1.0'::cube AS cube;
SELECT '-1.0'::cube AS cube;
Improve contrib/cube's handling of zero-D cubes, infinities, and NaNs. It's always been possible to create a zero-dimensional cube by converting from a zero-length float8 array, but cube_in failed to accept the '()' representation that cube_out produced for that case, resulting in a dump/reload hazard. Make it accept the case. Also fix a couple of other places that didn't behave sanely for zero-dimensional cubes: cube_size would produce 1.0 when surely the answer should be 0.0, and g_cube_distance risked a divide-by-zero failure. Likewise, it's always been possible to create cubes containing float8 infinity or NaN coordinate values, but cube_in couldn't parse such input, and cube_out produced platform-dependent spellings of the values. Convert them to use float8in_internal and float8out_internal so that the behavior will be the same as for float8, as we recently did for the core geometric types (cf commit 50861cd68). As in that commit, I don't pretend that this patch fixes all insane corner-case behaviors that may exist for NaNs, but it's a step forward. (This change allows removal of the separate cube_1.out and cube_3.out expected-files, as the platform dependency that previously required them is now gone: an underflowing coordinate value will now produce an error not plus or minus zero.) Make errors from cube_in follow project conventions as to spelling ("invalid input syntax for cube" not "bad cube representation") and errcode (INVALID_TEXT_REPRESENTATION not SYNTAX_ERROR). Also a few marginal code cleanups and comment improvements. Tom Lane, reviewed by Amul Sul Discussion: <15085.1472494782@sss.pgh.pa.us>
2016-09-27 17:38:33 +02:00
SELECT 'infinity'::cube AS cube;
SELECT '-infinity'::cube AS cube;
SELECT 'NaN'::cube AS cube;
The changes I have made are described in CHANGES. This was based on diffs to 7.3-devel and may not be applicable to 7.2. I have included a change covered by a previous bugfix patch I submitted (the problem with -.1 not being accepted by cube_in). It does not include a fix for the potential buffer overrun issue I reported for cube_yyerror in cubeparse.y. Bruno Wolff III
2002-08-30 01:03:58 +02:00
SELECT '.1234567890123456'::cube AS cube;
SELECT '+.1234567890123456'::cube AS cube;
SELECT '-.1234567890123456'::cube AS cube;
Gene Selkov's CUBE datatype (GiST example code)
2000-12-11 21:39:15 +01:00
-- simple lists (points)
Improve contrib/cube's handling of zero-D cubes, infinities, and NaNs. It's always been possible to create a zero-dimensional cube by converting from a zero-length float8 array, but cube_in failed to accept the '()' representation that cube_out produced for that case, resulting in a dump/reload hazard. Make it accept the case. Also fix a couple of other places that didn't behave sanely for zero-dimensional cubes: cube_size would produce 1.0 when surely the answer should be 0.0, and g_cube_distance risked a divide-by-zero failure. Likewise, it's always been possible to create cubes containing float8 infinity or NaN coordinate values, but cube_in couldn't parse such input, and cube_out produced platform-dependent spellings of the values. Convert them to use float8in_internal and float8out_internal so that the behavior will be the same as for float8, as we recently did for the core geometric types (cf commit 50861cd68). As in that commit, I don't pretend that this patch fixes all insane corner-case behaviors that may exist for NaNs, but it's a step forward. (This change allows removal of the separate cube_1.out and cube_3.out expected-files, as the platform dependency that previously required them is now gone: an underflowing coordinate value will now produce an error not plus or minus zero.) Make errors from cube_in follow project conventions as to spelling ("invalid input syntax for cube" not "bad cube representation") and errcode (INVALID_TEXT_REPRESENTATION not SYNTAX_ERROR). Also a few marginal code cleanups and comment improvements. Tom Lane, reviewed by Amul Sul Discussion: <15085.1472494782@sss.pgh.pa.us>
2016-09-27 17:38:33 +02:00
SELECT '()'::cube AS cube;
Gene Selkov's CUBE datatype (GiST example code)
2000-12-11 21:39:15 +01:00
SELECT '1,2'::cube AS cube;
SELECT '(1,2)'::cube AS cube;
SELECT '1,2,3,4,5'::cube AS cube;
SELECT '(1,2,3,4,5)'::cube AS cube;
-- double lists (cubes)
Improve contrib/cube's handling of zero-D cubes, infinities, and NaNs. It's always been possible to create a zero-dimensional cube by converting from a zero-length float8 array, but cube_in failed to accept the '()' representation that cube_out produced for that case, resulting in a dump/reload hazard. Make it accept the case. Also fix a couple of other places that didn't behave sanely for zero-dimensional cubes: cube_size would produce 1.0 when surely the answer should be 0.0, and g_cube_distance risked a divide-by-zero failure. Likewise, it's always been possible to create cubes containing float8 infinity or NaN coordinate values, but cube_in couldn't parse such input, and cube_out produced platform-dependent spellings of the values. Convert them to use float8in_internal and float8out_internal so that the behavior will be the same as for float8, as we recently did for the core geometric types (cf commit 50861cd68). As in that commit, I don't pretend that this patch fixes all insane corner-case behaviors that may exist for NaNs, but it's a step forward. (This change allows removal of the separate cube_1.out and cube_3.out expected-files, as the platform dependency that previously required them is now gone: an underflowing coordinate value will now produce an error not plus or minus zero.) Make errors from cube_in follow project conventions as to spelling ("invalid input syntax for cube" not "bad cube representation") and errcode (INVALID_TEXT_REPRESENTATION not SYNTAX_ERROR). Also a few marginal code cleanups and comment improvements. Tom Lane, reviewed by Amul Sul Discussion: <15085.1472494782@sss.pgh.pa.us>
2016-09-27 17:38:33 +02:00
SELECT '(),()'::cube AS cube;
Gene Selkov's CUBE datatype (GiST example code)
2000-12-11 21:39:15 +01:00
SELECT '(0),(0)'::cube AS cube;
SELECT '(0),(1)'::cube AS cube;
SELECT '[(0),(0)]'::cube AS cube;
SELECT '[(0),(1)]'::cube AS cube;
SELECT '(0,0,0,0),(0,0,0,0)'::cube AS cube;
SELECT '(0,0,0,0),(1,0,0,0)'::cube AS cube;
SELECT '[(0,0,0,0),(0,0,0,0)]'::cube AS cube;
SELECT '[(0,0,0,0),(1,0,0,0)]'::cube AS cube;
-- invalid input: parse errors
SELECT ''::cube AS cube;
SELECT 'ABC'::cube AS cube;
SELECT '[]'::cube AS cube;
SELECT '[()]'::cube AS cube;
SELECT '[(1)]'::cube AS cube;
SELECT '[(1),]'::cube AS cube;
SELECT '[(1),2]'::cube AS cube;
SELECT '[(1),(2),(3)]'::cube AS cube;
SELECT '1,'::cube AS cube;
SELECT '1,2,'::cube AS cube;
SELECT '1,,2'::cube AS cube;
SELECT '(1,)'::cube AS cube;
SELECT '(1,2,)'::cube AS cube;
SELECT '(1,,2)'::cube AS cube;
-- invalid input: semantic errors and trailing garbage
SELECT '[(1),(2)],'::cube AS cube; -- 0
SELECT '[(1,2,3),(2,3)]'::cube AS cube; -- 1
SELECT '[(1,2),(1,2,3)]'::cube AS cube; -- 1
SELECT '(1),(2),'::cube AS cube; -- 2
SELECT '(1,2,3),(2,3)'::cube AS cube; -- 3
SELECT '(1,2),(1,2,3)'::cube AS cube; -- 3
SELECT '(1,2,3)ab'::cube AS cube; -- 4
SELECT '(1,2,3)a'::cube AS cube; -- 5
SELECT '(1,2)('::cube AS cube; -- 5
SELECT '1,2ab'::cube AS cube; -- 6
SELECT '1 e7'::cube AS cube; -- 6
SELECT '1,2a'::cube AS cube; -- 7
SELECT '1..2'::cube AS cube; -- 7
Improve contrib/cube's handling of zero-D cubes, infinities, and NaNs. It's always been possible to create a zero-dimensional cube by converting from a zero-length float8 array, but cube_in failed to accept the '()' representation that cube_out produced for that case, resulting in a dump/reload hazard. Make it accept the case. Also fix a couple of other places that didn't behave sanely for zero-dimensional cubes: cube_size would produce 1.0 when surely the answer should be 0.0, and g_cube_distance risked a divide-by-zero failure. Likewise, it's always been possible to create cubes containing float8 infinity or NaN coordinate values, but cube_in couldn't parse such input, and cube_out produced platform-dependent spellings of the values. Convert them to use float8in_internal and float8out_internal so that the behavior will be the same as for float8, as we recently did for the core geometric types (cf commit 50861cd68). As in that commit, I don't pretend that this patch fixes all insane corner-case behaviors that may exist for NaNs, but it's a step forward. (This change allows removal of the separate cube_1.out and cube_3.out expected-files, as the platform dependency that previously required them is now gone: an underflowing coordinate value will now produce an error not plus or minus zero.) Make errors from cube_in follow project conventions as to spelling ("invalid input syntax for cube" not "bad cube representation") and errcode (INVALID_TEXT_REPRESENTATION not SYNTAX_ERROR). Also a few marginal code cleanups and comment improvements. Tom Lane, reviewed by Amul Sul Discussion: <15085.1472494782@sss.pgh.pa.us>
2016-09-27 17:38:33 +02:00
SELECT '-1e-700'::cube AS cube; -- out of range
Gene Selkov's CUBE datatype (GiST example code)
2000-12-11 21:39:15 +01:00
The attached patch provides cube with 4 functions for building cubes directly from float8 values. (As opposed to converting the values to strings and then parsing the strings.) The functions are: cube(float8) returns cube cube(float8,float8) returns cube cube(cube,float8) returns cube cube(cube,float8,float8) returns cube Bruno Wolff III
2003-02-13 06:26:50 +01:00
--
-- Testing building cubes from float8 values
--
SELECT cube(0::float8);
SELECT cube(1::float8);
SELECT cube(1,2);
SELECT cube(cube(1,2),3);
SELECT cube(cube(1,2),3,4);
SELECT cube(cube(cube(1,2),3,4),5);
SELECT cube(cube(cube(1,2),3,4),5,6);
--
-- Test that the text -> cube cast was installed.
--
SELECT '(0)'::text::cube;
/contrib/cube improvements: Update the calling convention for all external facing functions. By external facing, I mean all functions that are directly referenced in cube.sql. Prior to my update, all functions used the older V0 calling convention. They now use V1. New Functions: cube(float[]), which makes a zero volume cube from a float array cube(float[], float[]), which allows the user to create a cube from two float arrays; one for the upper right and one for the lower left coordinate. cube_subset(cube, int4[]), to allow you to reorder or choose a subset of dimensions from a cube, using index values specified in the array. Joshua Reich
2006-07-26 01:23:45 +02:00
--
-- Test the float[] -> cube cast
--
SELECT cube('{0,1,2}'::float[], '{3,4,5}'::float[]);
SELECT cube('{0,1,2}'::float[], '{3}'::float[]);
SELECT cube(NULL::float[], '{3}'::float[]);
SELECT cube('{0,1,2}'::float[]);
Remove useless whitespace at end of lines
2010-11-23 21:27:50 +01:00
SELECT cube_subset(cube('(1,3,5),(6,7,8)'), ARRAY[3,2,1,1]);
Extend cube on-disk format to pack points more tightly. If the lower left and upper right corners of a cube are the same, set a flag in the cube header, and only store one copy of the coordinates. That cuts the on-disk size into half for the common case that the cube datatype is used to represent points rather than boxes. The new format is backwards-compatible with the old one, so pg_upgrade still works. However, to get the space savings, the data needs to be rewritten. A simple VACUUM FULL or REINDEX is not enough, as the old Datums will just be moved to the new heap/index as is. A pg_dump and reload, or something similar like casting to text and back, will do the trick. This patch deliberately doesn't update all the alternative expected output files, as I don't have access to machines that produce those outputs. I'm not sure if they are still relevant, but if they are, the buildfarm will tell us and produce the diff required to fix it. If none of the buildfarm animals need them, they should be removed altogether. Patch by Stas Kelvich.
2013-10-21 19:59:42 +02:00
SELECT cube_subset(cube('(1,3,5),(1,3,5)'), ARRAY[3,2,1,1]);
Remove useless whitespace at end of lines
2010-11-23 21:27:50 +01:00
SELECT cube_subset(cube('(1,3,5),(6,7,8)'), ARRAY[4,0]);
Extend cube on-disk format to pack points more tightly. If the lower left and upper right corners of a cube are the same, set a flag in the cube header, and only store one copy of the coordinates. That cuts the on-disk size into half for the common case that the cube datatype is used to represent points rather than boxes. The new format is backwards-compatible with the old one, so pg_upgrade still works. However, to get the space savings, the data needs to be rewritten. A simple VACUUM FULL or REINDEX is not enough, as the old Datums will just be moved to the new heap/index as is. A pg_dump and reload, or something similar like casting to text and back, will do the trick. This patch deliberately doesn't update all the alternative expected output files, as I don't have access to machines that produce those outputs. I'm not sure if they are still relevant, but if they are, the buildfarm will tell us and produce the diff required to fix it. If none of the buildfarm animals need them, they should be removed altogether. Patch by Stas Kelvich.
2013-10-21 19:59:42 +02:00
SELECT cube_subset(cube('(6,7,8),(6,7,8)'), ARRAY[4,0]);
Enforce cube dimension limit in all cube construction functions contrib/cube has a limit to 100 dimensions for cube datatype. However, it's not enforced everywhere, and one can actually construct cube with more than 100 dimensions having then trouble with dump/restore. This commit add checks for dimensions limit in all functions responsible for cube construction. Backpatch to all supported versions. Reported-by: Andrew Gierth Discussion: https://postgr.es/m/87va7uybt4.fsf%40news-spur.riddles.org.uk Author: Andrey Borodin with small additions by me Review: Tom Lane Backpatch-through: 9.3
2018-08-30 13:18:53 +02:00
-- test for limits: this should pass
SELECT cube_subset(cube('(6,7,8),(6,7,8)'), array(SELECT 1 as a FROM generate_series(1,100)));
-- and this should fail
SELECT cube_subset(cube('(6,7,8),(6,7,8)'), array(SELECT 1 as a FROM generate_series(1,101)));
Extend cube on-disk format to pack points more tightly. If the lower left and upper right corners of a cube are the same, set a flag in the cube header, and only store one copy of the coordinates. That cuts the on-disk size into half for the common case that the cube datatype is used to represent points rather than boxes. The new format is backwards-compatible with the old one, so pg_upgrade still works. However, to get the space savings, the data needs to be rewritten. A simple VACUUM FULL or REINDEX is not enough, as the old Datums will just be moved to the new heap/index as is. A pg_dump and reload, or something similar like casting to text and back, will do the trick. This patch deliberately doesn't update all the alternative expected output files, as I don't have access to machines that produce those outputs. I'm not sure if they are still relevant, but if they are, the buildfarm will tell us and produce the diff required to fix it. If none of the buildfarm animals need them, they should be removed altogether. Patch by Stas Kelvich.
2013-10-21 19:59:42 +02:00
--
-- Test point processing
--
SELECT cube('(1,2),(1,2)'); -- cube_in
SELECT cube('{0,1,2}'::float[], '{0,1,2}'::float[]); -- cube_a_f8_f8
SELECT cube('{5,6,7,8}'::float[]); -- cube_a_f8
SELECT cube(1.37); -- cube_f8
SELECT cube(1.37, 1.37); -- cube_f8_f8
SELECT cube(cube(1,1), 42); -- cube_c_f8
SELECT cube(cube(1,2), 42); -- cube_c_f8
SELECT cube(cube(1,1), 42, 42); -- cube_c_f8_f8
SELECT cube(cube(1,1), 42, 24); -- cube_c_f8_f8
SELECT cube(cube(1,2), 42, 42); -- cube_c_f8_f8
SELECT cube(cube(1,2), 42, 24); -- cube_c_f8_f8
/contrib/cube improvements: Update the calling convention for all external facing functions. By external facing, I mean all functions that are directly referenced in cube.sql. Prior to my update, all functions used the older V0 calling convention. They now use V1. New Functions: cube(float[]), which makes a zero volume cube from a float array cube(float[], float[]), which allows the user to create a cube from two float arrays; one for the upper right and one for the lower left coordinate. cube_subset(cube, int4[]), to allow you to reorder or choose a subset of dimensions from a cube, using index values specified in the array. Joshua Reich
2006-07-26 01:23:45 +02:00
This is a comprehensive set of diffs (versus current CVS) that replaces those attached to the same message with the Earth Distance patches. Recent changes include changing the subscript in one place I forgot in the previous bugfix patch. A couple of added regression tests, which should help catch this mistake if it reappears. I also put in a limit of 100 dimensions in cube_large and cube_in to prevent making it easy to create very large cubes. Changing one define in cubedata.h will raise the limit if some needs more dimensions. Bruno Wolff III
2002-09-12 02:26:00 +02:00
--
-- Testing limit of CUBE_MAX_DIM dimensions check in cube_in.
--
Enforce cube dimension limit in all cube construction functions contrib/cube has a limit to 100 dimensions for cube datatype. However, it's not enforced everywhere, and one can actually construct cube with more than 100 dimensions having then trouble with dump/restore. This commit add checks for dimensions limit in all functions responsible for cube construction. Backpatch to all supported versions. Reported-by: Andrew Gierth Discussion: https://postgr.es/m/87va7uybt4.fsf%40news-spur.riddles.org.uk Author: Andrey Borodin with small additions by me Review: Tom Lane Backpatch-through: 9.3
2018-08-30 13:18:53 +02:00
-- create too big cube from literal
This is a comprehensive set of diffs (versus current CVS) that replaces those attached to the same message with the Earth Distance patches. Recent changes include changing the subscript in one place I forgot in the previous bugfix patch. A couple of added regression tests, which should help catch this mistake if it reappears. I also put in a limit of 100 dimensions in cube_large and cube_in to prevent making it easy to create very large cubes. Changing one define in cubedata.h will raise the limit if some needs more dimensions. Bruno Wolff III
2002-09-12 02:26:00 +02:00
select '(0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0)'::cube;
select '(0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0),(0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0)'::cube;
Enforce cube dimension limit in all cube construction functions contrib/cube has a limit to 100 dimensions for cube datatype. However, it's not enforced everywhere, and one can actually construct cube with more than 100 dimensions having then trouble with dump/restore. This commit add checks for dimensions limit in all functions responsible for cube construction. Backpatch to all supported versions. Reported-by: Andrew Gierth Discussion: https://postgr.es/m/87va7uybt4.fsf%40news-spur.riddles.org.uk Author: Andrey Borodin with small additions by me Review: Tom Lane Backpatch-through: 9.3
2018-08-30 13:18:53 +02:00
-- from an array
select cube(array(SELECT 0 as a FROM generate_series(1,101)));
select cube(array(SELECT 0 as a FROM generate_series(1,101)),array(SELECT 0 as a FROM generate_series(1,101)));
-- extend cube beyond limit
-- this should work
select cube(array(SELECT 0 as a FROM generate_series(1,100)));
select cube(array(SELECT 0 as a FROM generate_series(1,100)),array(SELECT 0 as a FROM generate_series(1,100)));
-- this should fail
select cube(cube(array(SELECT 0 as a FROM generate_series(1,100))), 0);
select cube(cube(array(SELECT 0 as a FROM generate_series(1,100)),array(SELECT 0 as a FROM generate_series(1,100))), 0, 0);
This is a comprehensive set of diffs (versus current CVS) that replaces those attached to the same message with the Earth Distance patches. Recent changes include changing the subscript in one place I forgot in the previous bugfix patch. A couple of added regression tests, which should help catch this mistake if it reappears. I also put in a limit of 100 dimensions in cube_large and cube_in to prevent making it easy to create very large cubes. Changing one define in cubedata.h will raise the limit if some needs more dimensions. Bruno Wolff III
2002-09-12 02:26:00 +02:00
Gene Selkov's CUBE datatype (GiST example code)
2000-12-11 21:39:15 +01:00
--
-- testing the operators
--
-- equality/inequality:
--
SELECT '24, 33.20'::cube = '24, 33.20'::cube AS bool;
SELECT '24, 33.20'::cube != '24, 33.20'::cube AS bool;
SELECT '24, 33.20'::cube = '24, 33.21'::cube AS bool;
SELECT '24, 33.20'::cube != '24, 33.21'::cube AS bool;
SELECT '(2,0),(3,1)'::cube = '(2,0,0,0,0),(3,1,0,0,0)'::cube AS bool;
SELECT '(2,0),(3,1)'::cube = '(2,0,0,0,0),(3,1,0,0,1)'::cube AS bool;
-- "lower than" / "greater than"
-- (these operators are not useful for anything but ordering)
--
SELECT '1'::cube > '2'::cube AS bool;
SELECT '1'::cube < '2'::cube AS bool;
SELECT '1,1'::cube > '1,2'::cube AS bool;
SELECT '1,1'::cube < '1,2'::cube AS bool;
SELECT '(2,0),(3,1)'::cube > '(2,0,0,0,0),(3,1,0,0,1)'::cube AS bool;
SELECT '(2,0),(3,1)'::cube < '(2,0,0,0,0),(3,1,0,0,1)'::cube AS bool;
SELECT '(2,0),(3,1)'::cube > '(2,0,0,0,1),(3,1,0,0,0)'::cube AS bool;
SELECT '(2,0),(3,1)'::cube < '(2,0,0,0,1),(3,1,0,0,0)'::cube AS bool;
SELECT '(2,0),(3,1)'::cube > '(2,0,0,0,0),(3,1,0,0,0)'::cube AS bool;
SELECT '(2,0),(3,1)'::cube < '(2,0,0,0,0),(3,1,0,0,0)'::cube AS bool;
SELECT '(2,0,0,0,0),(3,1,0,0,1)'::cube > '(2,0),(3,1)'::cube AS bool;
SELECT '(2,0,0,0,0),(3,1,0,0,1)'::cube < '(2,0),(3,1)'::cube AS bool;
SELECT '(2,0,0,0,1),(3,1,0,0,0)'::cube > '(2,0),(3,1)'::cube AS bool;
SELECT '(2,0,0,0,1),(3,1,0,0,0)'::cube < '(2,0),(3,1)'::cube AS bool;
SELECT '(2,0,0,0,0),(3,1,0,0,0)'::cube > '(2,0),(3,1)'::cube AS bool;
SELECT '(2,0,0,0,0),(3,1,0,0,0)'::cube < '(2,0),(3,1)'::cube AS bool;
-- "overlap"
--
SELECT '1'::cube && '1'::cube AS bool;
SELECT '1'::cube && '2'::cube AS bool;
SELECT '[(-1,-1,-1),(1,1,1)]'::cube && '0'::cube AS bool;
SELECT '[(-1,-1,-1),(1,1,1)]'::cube && '1'::cube AS bool;
SELECT '[(-1,-1,-1),(1,1,1)]'::cube && '1,1,1'::cube AS bool;
SELECT '[(-1,-1,-1),(1,1,1)]'::cube && '[(1,1,1),(2,2,2)]'::cube AS bool;
SELECT '[(-1,-1,-1),(1,1,1)]'::cube && '[(1,1),(2,2)]'::cube AS bool;
SELECT '[(-1,-1,-1),(1,1,1)]'::cube && '[(2,1,1),(2,2,2)]'::cube AS bool;
-- "contained in" (the left operand is the cube entirely enclosed by
-- the right operand):
--
Rename contrib contains/contained-by operators to @> and <@, per discussion.
2006-09-10 19:36:52 +02:00
SELECT '0'::cube <@ '0'::cube AS bool;
SELECT '0,0,0'::cube <@ '0,0,0'::cube AS bool;
SELECT '0,0'::cube <@ '0,0,1'::cube AS bool;
SELECT '0,0,0'::cube <@ '0,0,1'::cube AS bool;
SELECT '1,0,0'::cube <@ '0,0,1'::cube AS bool;
SELECT '(1,0,0),(0,0,1)'::cube <@ '(1,0,0),(0,0,1)'::cube AS bool;
SELECT '(1,0,0),(0,0,1)'::cube <@ '(-1,-1,-1),(1,1,1)'::cube AS bool;
SELECT '(1,0,0),(0,0,1)'::cube <@ '(-1,-1,-1,-1),(1,1,1,1)'::cube AS bool;
SELECT '0'::cube <@ '(-1),(1)'::cube AS bool;
SELECT '1'::cube <@ '(-1),(1)'::cube AS bool;
SELECT '-1'::cube <@ '(-1),(1)'::cube AS bool;
SELECT '(-1),(1)'::cube <@ '(-1),(1)'::cube AS bool;
SELECT '(-1),(1)'::cube <@ '(-1,-1),(1,1)'::cube AS bool;
SELECT '(-2),(1)'::cube <@ '(-1),(1)'::cube AS bool;
SELECT '(-2),(1)'::cube <@ '(-1,-1),(1,1)'::cube AS bool;
Gene Selkov's CUBE datatype (GiST example code)
2000-12-11 21:39:15 +01:00
-- "contains" (the left operand is the cube that entirely encloses the
-- right operand)
--
Rename contrib contains/contained-by operators to @> and <@, per discussion.
2006-09-10 19:36:52 +02:00
SELECT '0'::cube @> '0'::cube AS bool;
SELECT '0,0,0'::cube @> '0,0,0'::cube AS bool;
SELECT '0,0,1'::cube @> '0,0'::cube AS bool;
SELECT '0,0,1'::cube @> '0,0,0'::cube AS bool;
SELECT '0,0,1'::cube @> '1,0,0'::cube AS bool;
SELECT '(1,0,0),(0,0,1)'::cube @> '(1,0,0),(0,0,1)'::cube AS bool;
SELECT '(-1,-1,-1),(1,1,1)'::cube @> '(1,0,0),(0,0,1)'::cube AS bool;
SELECT '(-1,-1,-1,-1),(1,1,1,1)'::cube @> '(1,0,0),(0,0,1)'::cube AS bool;
SELECT '(-1),(1)'::cube @> '0'::cube AS bool;
SELECT '(-1),(1)'::cube @> '1'::cube AS bool;
SELECT '(-1),(1)'::cube @> '-1'::cube AS bool;
SELECT '(-1),(1)'::cube @> '(-1),(1)'::cube AS bool;
SELECT '(-1,-1),(1,1)'::cube @> '(-1),(1)'::cube AS bool;
SELECT '(-1),(1)'::cube @> '(-2),(1)'::cube AS bool;
SELECT '(-1,-1),(1,1)'::cube @> '(-2),(1)'::cube AS bool;
Gene Selkov's CUBE datatype (GiST example code)
2000-12-11 21:39:15 +01:00
The changes I have made are described in CHANGES. This was based on diffs to 7.3-devel and may not be applicable to 7.2. I have included a change covered by a previous bugfix patch I submitted (the problem with -.1 not being accepted by cube_in). It does not include a fix for the potential buffer overrun issue I reported for cube_yyerror in cubeparse.y. Bruno Wolff III
2002-08-30 01:03:58 +02:00
-- Test of distance function
--
SELECT cube_distance('(0)'::cube,'(2,2,2,2)'::cube);
SELECT cube_distance('(0)'::cube,'(.3,.4)'::cube);
Extend cube on-disk format to pack points more tightly. If the lower left and upper right corners of a cube are the same, set a flag in the cube header, and only store one copy of the coordinates. That cuts the on-disk size into half for the common case that the cube datatype is used to represent points rather than boxes. The new format is backwards-compatible with the old one, so pg_upgrade still works. However, to get the space savings, the data needs to be rewritten. A simple VACUUM FULL or REINDEX is not enough, as the old Datums will just be moved to the new heap/index as is. A pg_dump and reload, or something similar like casting to text and back, will do the trick. This patch deliberately doesn't update all the alternative expected output files, as I don't have access to machines that produce those outputs. I'm not sure if they are still relevant, but if they are, the buildfarm will tell us and produce the diff required to fix it. If none of the buildfarm animals need them, they should be removed altogether. Patch by Stas Kelvich.
2013-10-21 19:59:42 +02:00
SELECT cube_distance('(2,3,4)'::cube,'(2,3,4)'::cube);
SELECT cube_distance('(42,42,42,42)'::cube,'(137,137,137,137)'::cube);
SELECT cube_distance('(42,42,42)'::cube,'(137,137)'::cube);
The changes I have made are described in CHANGES. This was based on diffs to 7.3-devel and may not be applicable to 7.2. I have included a change covered by a previous bugfix patch I submitted (the problem with -.1 not being accepted by cube_in). It does not include a fix for the potential buffer overrun issue I reported for cube_yyerror in cubeparse.y. Bruno Wolff III
2002-08-30 01:03:58 +02:00
-- Test of cube function (text to cube)
--
Downgrade implicit casts to text to be assignment-only, except for the ones from the other string-category types; this eliminates a lot of surprising interpretations that the parser could formerly make when there was no directly applicable operator. Create a general mechanism that supports casts to and from the standard string types (text,varchar,bpchar) for *every* datatype, by invoking the datatype's I/O functions. These new casts are assignment-only in the to-string direction, explicit-only in the other, and therefore should create no surprising behavior. Remove a bunch of thereby-obsoleted datatype-specific casting functions. The "general mechanism" is a new expression node type CoerceViaIO that can actually convert between *any* two datatypes if their external text representations are compatible. This is more general than needed for the immediate feature, but might be useful in plpgsql or other places in future. This commit does nothing about the issue that applying the concatenation operator || to non-text types will now fail, often with strange error messages due to misinterpreting the operator as array concatenation. Since it often (not always) worked before, we should either make it succeed or at least give a more user-friendly error; but details are still under debate. Peter Eisentraut and Tom Lane
2007-06-05 23:31:09 +02:00
SELECT cube('(1,1.2)'::text);
The changes I have made are described in CHANGES. This was based on diffs to 7.3-devel and may not be applicable to 7.2. I have included a change covered by a previous bugfix patch I submitted (the problem with -.1 not being accepted by cube_in). It does not include a fix for the potential buffer overrun issue I reported for cube_yyerror in cubeparse.y. Bruno Wolff III
2002-08-30 01:03:58 +02:00
SELECT cube(NULL);
-- Test of cube_dim function (dimensions stored in cube)
--
SELECT cube_dim('(0)'::cube);
SELECT cube_dim('(0,0)'::cube);
SELECT cube_dim('(0,0,0)'::cube);
Extend cube on-disk format to pack points more tightly. If the lower left and upper right corners of a cube are the same, set a flag in the cube header, and only store one copy of the coordinates. That cuts the on-disk size into half for the common case that the cube datatype is used to represent points rather than boxes. The new format is backwards-compatible with the old one, so pg_upgrade still works. However, to get the space savings, the data needs to be rewritten. A simple VACUUM FULL or REINDEX is not enough, as the old Datums will just be moved to the new heap/index as is. A pg_dump and reload, or something similar like casting to text and back, will do the trick. This patch deliberately doesn't update all the alternative expected output files, as I don't have access to machines that produce those outputs. I'm not sure if they are still relevant, but if they are, the buildfarm will tell us and produce the diff required to fix it. If none of the buildfarm animals need them, they should be removed altogether. Patch by Stas Kelvich.
2013-10-21 19:59:42 +02:00
SELECT cube_dim('(42,42,42),(42,42,42)'::cube);
SELECT cube_dim('(4,8,15,16,23),(4,8,15,16,23)'::cube);
The changes I have made are described in CHANGES. This was based on diffs to 7.3-devel and may not be applicable to 7.2. I have included a change covered by a previous bugfix patch I submitted (the problem with -.1 not being accepted by cube_in). It does not include a fix for the potential buffer overrun issue I reported for cube_yyerror in cubeparse.y. Bruno Wolff III
2002-08-30 01:03:58 +02:00
Fix typos in comments. Backpatch to all supported versions, where applicable, to make backpatching of future fixes go more smoothly. Josh Soref Discussion: https://www.postgresql.org/message-id/CACZqfqCf+5qRztLPgmmosr-B0Ye4srWzzw_mo4c_8_B_mtjmJQ@mail.gmail.com
2017-02-06 10:33:58 +01:00
-- Test of cube_ll_coord function (retrieves LL coordinate values)
The changes I have made are described in CHANGES. This was based on diffs to 7.3-devel and may not be applicable to 7.2. I have included a change covered by a previous bugfix patch I submitted (the problem with -.1 not being accepted by cube_in). It does not include a fix for the potential buffer overrun issue I reported for cube_yyerror in cubeparse.y. Bruno Wolff III
2002-08-30 01:03:58 +02:00
--
SELECT cube_ll_coord('(-1,1),(2,-2)'::cube, 1);
SELECT cube_ll_coord('(-1,1),(2,-2)'::cube, 2);
SELECT cube_ll_coord('(-1,1),(2,-2)'::cube, 3);
Extend cube on-disk format to pack points more tightly. If the lower left and upper right corners of a cube are the same, set a flag in the cube header, and only store one copy of the coordinates. That cuts the on-disk size into half for the common case that the cube datatype is used to represent points rather than boxes. The new format is backwards-compatible with the old one, so pg_upgrade still works. However, to get the space savings, the data needs to be rewritten. A simple VACUUM FULL or REINDEX is not enough, as the old Datums will just be moved to the new heap/index as is. A pg_dump and reload, or something similar like casting to text and back, will do the trick. This patch deliberately doesn't update all the alternative expected output files, as I don't have access to machines that produce those outputs. I'm not sure if they are still relevant, but if they are, the buildfarm will tell us and produce the diff required to fix it. If none of the buildfarm animals need them, they should be removed altogether. Patch by Stas Kelvich.
2013-10-21 19:59:42 +02:00
SELECT cube_ll_coord('(1,2),(1,2)'::cube, 1);
SELECT cube_ll_coord('(1,2),(1,2)'::cube, 2);
SELECT cube_ll_coord('(1,2),(1,2)'::cube, 3);
SELECT cube_ll_coord('(42,137)'::cube, 1);
SELECT cube_ll_coord('(42,137)'::cube, 2);
SELECT cube_ll_coord('(42,137)'::cube, 3);
The changes I have made are described in CHANGES. This was based on diffs to 7.3-devel and may not be applicable to 7.2. I have included a change covered by a previous bugfix patch I submitted (the problem with -.1 not being accepted by cube_in). It does not include a fix for the potential buffer overrun issue I reported for cube_yyerror in cubeparse.y. Bruno Wolff III
2002-08-30 01:03:58 +02:00
Fix typos in comments. Backpatch to all supported versions, where applicable, to make backpatching of future fixes go more smoothly. Josh Soref Discussion: https://www.postgresql.org/message-id/CACZqfqCf+5qRztLPgmmosr-B0Ye4srWzzw_mo4c_8_B_mtjmJQ@mail.gmail.com
2017-02-06 10:33:58 +01:00
-- Test of cube_ur_coord function (retrieves UR coordinate values)
The changes I have made are described in CHANGES. This was based on diffs to 7.3-devel and may not be applicable to 7.2. I have included a change covered by a previous bugfix patch I submitted (the problem with -.1 not being accepted by cube_in). It does not include a fix for the potential buffer overrun issue I reported for cube_yyerror in cubeparse.y. Bruno Wolff III
2002-08-30 01:03:58 +02:00
--
SELECT cube_ur_coord('(-1,1),(2,-2)'::cube, 1);
SELECT cube_ur_coord('(-1,1),(2,-2)'::cube, 2);
SELECT cube_ur_coord('(-1,1),(2,-2)'::cube, 3);
Extend cube on-disk format to pack points more tightly. If the lower left and upper right corners of a cube are the same, set a flag in the cube header, and only store one copy of the coordinates. That cuts the on-disk size into half for the common case that the cube datatype is used to represent points rather than boxes. The new format is backwards-compatible with the old one, so pg_upgrade still works. However, to get the space savings, the data needs to be rewritten. A simple VACUUM FULL or REINDEX is not enough, as the old Datums will just be moved to the new heap/index as is. A pg_dump and reload, or something similar like casting to text and back, will do the trick. This patch deliberately doesn't update all the alternative expected output files, as I don't have access to machines that produce those outputs. I'm not sure if they are still relevant, but if they are, the buildfarm will tell us and produce the diff required to fix it. If none of the buildfarm animals need them, they should be removed altogether. Patch by Stas Kelvich.
2013-10-21 19:59:42 +02:00
SELECT cube_ur_coord('(1,2),(1,2)'::cube, 1);
SELECT cube_ur_coord('(1,2),(1,2)'::cube, 2);
SELECT cube_ur_coord('(1,2),(1,2)'::cube, 3);
SELECT cube_ur_coord('(42,137)'::cube, 1);
SELECT cube_ur_coord('(42,137)'::cube, 2);
SELECT cube_ur_coord('(42,137)'::cube, 3);
The changes I have made are described in CHANGES. This was based on diffs to 7.3-devel and may not be applicable to 7.2. I have included a change covered by a previous bugfix patch I submitted (the problem with -.1 not being accepted by cube_in). It does not include a fix for the potential buffer overrun issue I reported for cube_yyerror in cubeparse.y. Bruno Wolff III
2002-08-30 01:03:58 +02:00
-- Test of cube_is_point
--
SELECT cube_is_point('(0)'::cube);
SELECT cube_is_point('(0,1,2)'::cube);
SELECT cube_is_point('(0,1,2),(0,1,2)'::cube);
SELECT cube_is_point('(0,1,2),(-1,1,2)'::cube);
SELECT cube_is_point('(0,1,2),(0,-1,2)'::cube);
SELECT cube_is_point('(0,1,2),(0,1,-2)'::cube);
-- Test of cube_enlarge (enlarging and shrinking cubes)
--
SELECT cube_enlarge('(0)'::cube, 0, 0);
SELECT cube_enlarge('(0)'::cube, 0, 1);
SELECT cube_enlarge('(0)'::cube, 0, 2);
This is a comprehensive set of diffs (versus current CVS) that replaces those attached to the same message with the Earth Distance patches. Recent changes include changing the subscript in one place I forgot in the previous bugfix patch. A couple of added regression tests, which should help catch this mistake if it reappears. I also put in a limit of 100 dimensions in cube_large and cube_in to prevent making it easy to create very large cubes. Changing one define in cubedata.h will raise the limit if some needs more dimensions. Bruno Wolff III
2002-09-12 02:26:00 +02:00
SELECT cube_enlarge('(2),(-2)'::cube, 0, 4);
The changes I have made are described in CHANGES. This was based on diffs to 7.3-devel and may not be applicable to 7.2. I have included a change covered by a previous bugfix patch I submitted (the problem with -.1 not being accepted by cube_in). It does not include a fix for the potential buffer overrun issue I reported for cube_yyerror in cubeparse.y. Bruno Wolff III
2002-08-30 01:03:58 +02:00
SELECT cube_enlarge('(0)'::cube, 1, 0);
SELECT cube_enlarge('(0)'::cube, 1, 1);
SELECT cube_enlarge('(0)'::cube, 1, 2);
This is a comprehensive set of diffs (versus current CVS) that replaces those attached to the same message with the Earth Distance patches. Recent changes include changing the subscript in one place I forgot in the previous bugfix patch. A couple of added regression tests, which should help catch this mistake if it reappears. I also put in a limit of 100 dimensions in cube_large and cube_in to prevent making it easy to create very large cubes. Changing one define in cubedata.h will raise the limit if some needs more dimensions. Bruno Wolff III
2002-09-12 02:26:00 +02:00
SELECT cube_enlarge('(2),(-2)'::cube, 1, 4);
The changes I have made are described in CHANGES. This was based on diffs to 7.3-devel and may not be applicable to 7.2. I have included a change covered by a previous bugfix patch I submitted (the problem with -.1 not being accepted by cube_in). It does not include a fix for the potential buffer overrun issue I reported for cube_yyerror in cubeparse.y. Bruno Wolff III
2002-08-30 01:03:58 +02:00
SELECT cube_enlarge('(0)'::cube, -1, 0);
SELECT cube_enlarge('(0)'::cube, -1, 1);
SELECT cube_enlarge('(0)'::cube, -1, 2);
This is a comprehensive set of diffs (versus current CVS) that replaces those attached to the same message with the Earth Distance patches. Recent changes include changing the subscript in one place I forgot in the previous bugfix patch. A couple of added regression tests, which should help catch this mistake if it reappears. I also put in a limit of 100 dimensions in cube_large and cube_in to prevent making it easy to create very large cubes. Changing one define in cubedata.h will raise the limit if some needs more dimensions. Bruno Wolff III
2002-09-12 02:26:00 +02:00
SELECT cube_enlarge('(2),(-2)'::cube, -1, 4);
The changes I have made are described in CHANGES. This was based on diffs to 7.3-devel and may not be applicable to 7.2. I have included a change covered by a previous bugfix patch I submitted (the problem with -.1 not being accepted by cube_in). It does not include a fix for the potential buffer overrun issue I reported for cube_yyerror in cubeparse.y. Bruno Wolff III
2002-08-30 01:03:58 +02:00
SELECT cube_enlarge('(0,0,0)'::cube, 1, 0);
SELECT cube_enlarge('(0,0,0)'::cube, 1, 2);
SELECT cube_enlarge('(2,-2),(-3,7)'::cube, 1, 2);
SELECT cube_enlarge('(2,-2),(-3,7)'::cube, 3, 2);
SELECT cube_enlarge('(2,-2),(-3,7)'::cube, -1, 2);
SELECT cube_enlarge('(2,-2),(-3,7)'::cube, -3, 2);
Extend cube on-disk format to pack points more tightly. If the lower left and upper right corners of a cube are the same, set a flag in the cube header, and only store one copy of the coordinates. That cuts the on-disk size into half for the common case that the cube datatype is used to represent points rather than boxes. The new format is backwards-compatible with the old one, so pg_upgrade still works. However, to get the space savings, the data needs to be rewritten. A simple VACUUM FULL or REINDEX is not enough, as the old Datums will just be moved to the new heap/index as is. A pg_dump and reload, or something similar like casting to text and back, will do the trick. This patch deliberately doesn't update all the alternative expected output files, as I don't have access to machines that produce those outputs. I'm not sure if they are still relevant, but if they are, the buildfarm will tell us and produce the diff required to fix it. If none of the buildfarm animals need them, they should be removed altogether. Patch by Stas Kelvich.
2013-10-21 19:59:42 +02:00
SELECT cube_enlarge('(42,-23,-23),(42,23,23)'::cube, -23, 5);
SELECT cube_enlarge('(42,-23,-23),(42,23,23)'::cube, -24, 5);
-- Test of cube_union (MBR for two cubes)
--
SELECT cube_union('(1,2),(3,4)'::cube, '(5,6,7),(8,9,10)'::cube);
SELECT cube_union('(1,2)'::cube, '(4,2,0,0)'::cube);
SELECT cube_union('(1,2),(1,2)'::cube, '(4,2),(4,2)'::cube);
SELECT cube_union('(1,2),(1,2)'::cube, '(1,2),(1,2)'::cube);
SELECT cube_union('(1,2),(1,2)'::cube, '(1,2,0),(1,2,0)'::cube);
-- Test of cube_inter
--
SELECT cube_inter('(1,2),(10,11)'::cube, '(3,4), (16,15)'::cube); -- intersects
SELECT cube_inter('(1,2),(10,11)'::cube, '(3,4), (6,5)'::cube); -- includes
SELECT cube_inter('(1,2),(10,11)'::cube, '(13,14), (16,15)'::cube); -- no intersection
SELECT cube_inter('(1,2),(10,11)'::cube, '(3,14), (16,15)'::cube); -- no intersection, but one dimension intersects
SELECT cube_inter('(1,2),(10,11)'::cube, '(10,11), (16,15)'::cube); -- point intersection
SELECT cube_inter('(1,2,3)'::cube, '(1,2,3)'::cube); -- point args
SELECT cube_inter('(1,2,3)'::cube, '(5,6,3)'::cube); -- point args
-- Test of cube_size
--
SELECT cube_size('(4,8),(15,16)'::cube);
SELECT cube_size('(42,137)'::cube);
Gene Selkov's CUBE datatype (GiST example code)
2000-12-11 21:39:15 +01:00
Change floating-point output format for improved performance. Previously, floating-point output was done by rounding to a specific decimal precision; by default, to 6 or 15 decimal digits (losing information) or as requested using extra_float_digits. Drivers that wanted exact float values, and applications like pg_dump that must preserve values exactly, set extra_float_digits=3 (or sometimes 2 for historical reasons, though this isn't enough for float4). Unfortunately, decimal rounded output is slow enough to become a noticable bottleneck when dealing with large result sets or COPY of large tables when many floating-point values are involved. Floating-point output can be done much faster when the output is not rounded to a specific decimal length, but rather is chosen as the shortest decimal representation that is closer to the original float value than to any other value representable in the same precision. The recently published Ryu algorithm by Ulf Adams is both relatively simple and remarkably fast. Accordingly, change float4out/float8out to output shortest decimal representations if extra_float_digits is greater than 0, and make that the new default. Applications that need rounded output can set extra_float_digits back to 0 or below, and take the resulting performance hit. We make one concession to portability for systems with buggy floating-point input: we do not output decimal values that fall exactly halfway between adjacent representable binary values (which would rely on the reader doing round-to-nearest-even correctly). This is known to be a problem at least for VS2013 on Windows. Our version of the Ryu code originates from https://github.com/ulfjack/ryu/ at commit c9c3fb1979, but with the following (significant) modifications: - Output format is changed to use fixed-point notation for small exponents, as printf would, and also to use lowercase 'e', a minimum of 2 exponent digits, and a mandatory sign on the exponent, to keep the formatting as close as possible to previous output. - The output of exact midpoint values is disabled as noted above. - The integer fast-path code is changed somewhat (since we have fixed-point output and the upstream did not). - Our project style has been largely applied to the code with the exception of C99 declaration-after-statement, which has been retained as an exception to our present policy. - Most of upstream's debugging and conditionals are removed, and we use our own configure tests to determine things like uint128 availability. Changing the float output format obviously affects a number of regression tests. This patch uses an explicit setting of extra_float_digits=0 for test output that is not expected to be exactly reproducible (e.g. due to numerical instability or differing algorithms for transcendental functions). Conversions from floats to numeric are unchanged by this patch. These may appear in index expressions and it is not yet clear whether any change should be made, so that can be left for another day. This patch assumes that the only supported floating point format is now IEEE format, and the documentation is updated to reflect that. Code by me, adapting the work of Ulf Adams and other contributors. References: https://dl.acm.org/citation.cfm?id=3192369 Reviewed-by: Tom Lane, Andres Freund, Donald Dong Discussion: https://postgr.es/m/87r2el1bx6.fsf@news-spur.riddles.org.uk
2019-02-13 16:20:33 +01:00
-- Test of distances (euclidean distance may not be bit-exact)
Cube extension kNN support Introduce distance operators over cubes: <#> taxicab distance <-> euclidean distance <=> chebyshev distance Also add kNN support of those distances in GiST opclass. Author: Stas Kelvich
2015-12-18 12:38:27 +01:00
--
Change floating-point output format for improved performance. Previously, floating-point output was done by rounding to a specific decimal precision; by default, to 6 or 15 decimal digits (losing information) or as requested using extra_float_digits. Drivers that wanted exact float values, and applications like pg_dump that must preserve values exactly, set extra_float_digits=3 (or sometimes 2 for historical reasons, though this isn't enough for float4). Unfortunately, decimal rounded output is slow enough to become a noticable bottleneck when dealing with large result sets or COPY of large tables when many floating-point values are involved. Floating-point output can be done much faster when the output is not rounded to a specific decimal length, but rather is chosen as the shortest decimal representation that is closer to the original float value than to any other value representable in the same precision. The recently published Ryu algorithm by Ulf Adams is both relatively simple and remarkably fast. Accordingly, change float4out/float8out to output shortest decimal representations if extra_float_digits is greater than 0, and make that the new default. Applications that need rounded output can set extra_float_digits back to 0 or below, and take the resulting performance hit. We make one concession to portability for systems with buggy floating-point input: we do not output decimal values that fall exactly halfway between adjacent representable binary values (which would rely on the reader doing round-to-nearest-even correctly). This is known to be a problem at least for VS2013 on Windows. Our version of the Ryu code originates from https://github.com/ulfjack/ryu/ at commit c9c3fb1979, but with the following (significant) modifications: - Output format is changed to use fixed-point notation for small exponents, as printf would, and also to use lowercase 'e', a minimum of 2 exponent digits, and a mandatory sign on the exponent, to keep the formatting as close as possible to previous output. - The output of exact midpoint values is disabled as noted above. - The integer fast-path code is changed somewhat (since we have fixed-point output and the upstream did not). - Our project style has been largely applied to the code with the exception of C99 declaration-after-statement, which has been retained as an exception to our present policy. - Most of upstream's debugging and conditionals are removed, and we use our own configure tests to determine things like uint128 availability. Changing the float output format obviously affects a number of regression tests. This patch uses an explicit setting of extra_float_digits=0 for test output that is not expected to be exactly reproducible (e.g. due to numerical instability or differing algorithms for transcendental functions). Conversions from floats to numeric are unchanged by this patch. These may appear in index expressions and it is not yet clear whether any change should be made, so that can be left for another day. This patch assumes that the only supported floating point format is now IEEE format, and the documentation is updated to reflect that. Code by me, adapting the work of Ulf Adams and other contributors. References: https://dl.acm.org/citation.cfm?id=3192369 Reviewed-by: Tom Lane, Andres Freund, Donald Dong Discussion: https://postgr.es/m/87r2el1bx6.fsf@news-spur.riddles.org.uk
2019-02-13 16:20:33 +01:00
SET extra_float_digits = 0;
Cube extension kNN support Introduce distance operators over cubes: <#> taxicab distance <-> euclidean distance <=> chebyshev distance Also add kNN support of those distances in GiST opclass. Author: Stas Kelvich
2015-12-18 12:38:27 +01:00
SELECT cube_distance('(1,1)'::cube, '(4,5)'::cube);
SELECT '(1,1)'::cube <-> '(4,5)'::cube as d_e;
Change floating-point output format for improved performance. Previously, floating-point output was done by rounding to a specific decimal precision; by default, to 6 or 15 decimal digits (losing information) or as requested using extra_float_digits. Drivers that wanted exact float values, and applications like pg_dump that must preserve values exactly, set extra_float_digits=3 (or sometimes 2 for historical reasons, though this isn't enough for float4). Unfortunately, decimal rounded output is slow enough to become a noticable bottleneck when dealing with large result sets or COPY of large tables when many floating-point values are involved. Floating-point output can be done much faster when the output is not rounded to a specific decimal length, but rather is chosen as the shortest decimal representation that is closer to the original float value than to any other value representable in the same precision. The recently published Ryu algorithm by Ulf Adams is both relatively simple and remarkably fast. Accordingly, change float4out/float8out to output shortest decimal representations if extra_float_digits is greater than 0, and make that the new default. Applications that need rounded output can set extra_float_digits back to 0 or below, and take the resulting performance hit. We make one concession to portability for systems with buggy floating-point input: we do not output decimal values that fall exactly halfway between adjacent representable binary values (which would rely on the reader doing round-to-nearest-even correctly). This is known to be a problem at least for VS2013 on Windows. Our version of the Ryu code originates from https://github.com/ulfjack/ryu/ at commit c9c3fb1979, but with the following (significant) modifications: - Output format is changed to use fixed-point notation for small exponents, as printf would, and also to use lowercase 'e', a minimum of 2 exponent digits, and a mandatory sign on the exponent, to keep the formatting as close as possible to previous output. - The output of exact midpoint values is disabled as noted above. - The integer fast-path code is changed somewhat (since we have fixed-point output and the upstream did not). - Our project style has been largely applied to the code with the exception of C99 declaration-after-statement, which has been retained as an exception to our present policy. - Most of upstream's debugging and conditionals are removed, and we use our own configure tests to determine things like uint128 availability. Changing the float output format obviously affects a number of regression tests. This patch uses an explicit setting of extra_float_digits=0 for test output that is not expected to be exactly reproducible (e.g. due to numerical instability or differing algorithms for transcendental functions). Conversions from floats to numeric are unchanged by this patch. These may appear in index expressions and it is not yet clear whether any change should be made, so that can be left for another day. This patch assumes that the only supported floating point format is now IEEE format, and the documentation is updated to reflect that. Code by me, adapting the work of Ulf Adams and other contributors. References: https://dl.acm.org/citation.cfm?id=3192369 Reviewed-by: Tom Lane, Andres Freund, Donald Dong Discussion: https://postgr.es/m/87r2el1bx6.fsf@news-spur.riddles.org.uk
2019-02-13 16:20:33 +01:00
RESET extra_float_digits;
Cube extension kNN support Introduce distance operators over cubes: <#> taxicab distance <-> euclidean distance <=> chebyshev distance Also add kNN support of those distances in GiST opclass. Author: Stas Kelvich
2015-12-18 12:38:27 +01:00
SELECT distance_chebyshev('(1,1)'::cube, '(4,5)'::cube);
SELECT '(1,1)'::cube <=> '(4,5)'::cube as d_c;
SELECT distance_taxicab('(1,1)'::cube, '(4,5)'::cube);
SELECT '(1,1)'::cube <#> '(4,5)'::cube as d_t;
-- zero for overlapping
SELECT cube_distance('(2,2),(10,10)'::cube, '(0,0),(5,5)'::cube);
SELECT distance_chebyshev('(2,2),(10,10)'::cube, '(0,0),(5,5)'::cube);
SELECT distance_taxicab('(2,2),(10,10)'::cube, '(0,0),(5,5)'::cube);
-- coordinate access
SELECT cube(array[10,20,30], array[40,50,60])->1;
SELECT cube(array[40,50,60], array[10,20,30])->1;
SELECT cube(array[10,20,30], array[40,50,60])->6;
SELECT cube(array[10,20,30], array[40,50,60])->0;
SELECT cube(array[10,20,30], array[40,50,60])->7;
SELECT cube(array[10,20,30], array[40,50,60])->-1;
SELECT cube(array[10,20,30], array[40,50,60])->-6;
SELECT cube(array[10,20,30])->3;
SELECT cube(array[10,20,30])->6;
SELECT cube(array[10,20,30])->-6;
-- "normalized" coordinate access
SELECT cube(array[10,20,30], array[40,50,60])~>1;
SELECT cube(array[40,50,60], array[10,20,30])~>1;
SELECT cube(array[10,20,30], array[40,50,60])~>2;
SELECT cube(array[40,50,60], array[10,20,30])~>2;
SELECT cube(array[10,20,30], array[40,50,60])~>3;
SELECT cube(array[40,50,60], array[10,20,30])~>3;
SELECT cube(array[40,50,60], array[10,20,30])~>0;
SELECT cube(array[40,50,60], array[10,20,30])~>4;
SELECT cube(array[40,50,60], array[10,20,30])~>(-1);
Gene Selkov's CUBE datatype (GiST example code)
2000-12-11 21:39:15 +01:00
-- Load some example data and build the index
Remove useless whitespace at end of lines
2010-11-23 21:27:50 +01:00
--
Gene Selkov's CUBE datatype (GiST example code)
2000-12-11 21:39:15 +01:00
CREATE TABLE test_cube (c cube);
\copy test_cube from 'data/test_cube.data'
CREATE INDEX test_cube_ix ON test_cube USING gist (c);
Remove useless whitespace at end of lines
2010-11-23 21:27:50 +01:00
SELECT * FROM test_cube WHERE c && '(3000,1000),(0,0)' ORDER BY c;
Gene Selkov's CUBE datatype (GiST example code)
2000-12-11 21:39:15 +01:00
Remove useless whitespace at end of lines
2010-11-23 21:27:50 +01:00
-- Test sorting
SELECT * FROM test_cube WHERE c && '(3000,1000),(0,0)' GROUP BY c ORDER BY c;
Cube extension kNN support Introduce distance operators over cubes: <#> taxicab distance <-> euclidean distance <=> chebyshev distance Also add kNN support of those distances in GiST opclass. Author: Stas Kelvich
2015-12-18 12:38:27 +01:00
Support index-only scans in contrib/cube and contrib/seg GiST indexes. To do this, we only have to remove the compress and decompress support functions, which have never done anything more than detoasting. In the wake of commit d3a4f89d8, this results in automatically enabling index-only scans, since the core code will now know that the stored representation is the same as the original data (up to detoasting). The only exciting part of this is that ALTER OPERATOR FAMILY lacks a way to drop a support function that was declared as being part of an opclass rather than being loose in the family. For the moment, we'll hack our way to a solution with a manual update of the pg_depend entry type, which is what distinguishes the two cases. Perhaps someday it'll be worth providing a cleaner way to do that, but for now it seems like a very niche problem. Note that the underlying C functions remain, to support use of the shared libraries with older versions of the modules' SQL declarations. Someday we may be able to remove them, but not soon. Andrey Borodin, reviewed by me Discussion: https://postgr.es/m/D0F53A05-4F4A-4DEC-8339-3C069FA0EE11@yandex-team.ru
2017-11-21 02:25:18 +01:00
-- Test index-only scans
SET enable_bitmapscan = false;
EXPLAIN (COSTS OFF)
SELECT c FROM test_cube WHERE c <@ '(3000,1000),(0,0)' ORDER BY c;
SELECT c FROM test_cube WHERE c <@ '(3000,1000),(0,0)' ORDER BY c;
RESET enable_bitmapscan;
Fix behavior of ~> (cube, int) operator ~> (cube, int) operator was especially designed for knn-gist search. However, it appears that knn-gist search can't work correctly with current behavior of this operator when dataset contains cubes of variable dimensionality. In this case, the same value of second operator argument can point to different dimension depending on dimensionality of particular cube. Such behavior is incompatible with gist indexing of cubes, and knn-gist doesn't work correctly for it. This patch changes behavior of ~> (cube, int) operator by introducing dimension numbering where value of second argument unambiguously identifies number of dimension. With new behavior, this operator can be correctly supported by knn-gist. Relevant changes to cube operator class are also included. Backpatch to v9.6 where operator was introduced. Since behavior of ~> (cube, int) operator is changed, depending entities must be refreshed after upgrade. Such as, expression indexes using this operator must be reindexed, materialized views must be rebuilt, stored procedures and client code must be revised to correctly use new behavior. That should be mentioned in release notes. Noticed by: Tomas Vondra Author: Alexander Korotkov Reviewed by: Tomas Vondra, Andrey Borodin Discussion: https://www.postgresql.org/message-id/flat/a9657f6a-b497-36ff-e56-482a2c7e3292@2ndquadrant.com
2018-01-11 12:41:14 +01:00
-- Test kNN
INSERT INTO test_cube VALUES ('(1,1)'), ('(100000)'), ('(0, 100000)'); -- Some corner cases
SET enable_seqscan = false;
-- Test different metrics
Change floating-point output format for improved performance. Previously, floating-point output was done by rounding to a specific decimal precision; by default, to 6 or 15 decimal digits (losing information) or as requested using extra_float_digits. Drivers that wanted exact float values, and applications like pg_dump that must preserve values exactly, set extra_float_digits=3 (or sometimes 2 for historical reasons, though this isn't enough for float4). Unfortunately, decimal rounded output is slow enough to become a noticable bottleneck when dealing with large result sets or COPY of large tables when many floating-point values are involved. Floating-point output can be done much faster when the output is not rounded to a specific decimal length, but rather is chosen as the shortest decimal representation that is closer to the original float value than to any other value representable in the same precision. The recently published Ryu algorithm by Ulf Adams is both relatively simple and remarkably fast. Accordingly, change float4out/float8out to output shortest decimal representations if extra_float_digits is greater than 0, and make that the new default. Applications that need rounded output can set extra_float_digits back to 0 or below, and take the resulting performance hit. We make one concession to portability for systems with buggy floating-point input: we do not output decimal values that fall exactly halfway between adjacent representable binary values (which would rely on the reader doing round-to-nearest-even correctly). This is known to be a problem at least for VS2013 on Windows. Our version of the Ryu code originates from https://github.com/ulfjack/ryu/ at commit c9c3fb1979, but with the following (significant) modifications: - Output format is changed to use fixed-point notation for small exponents, as printf would, and also to use lowercase 'e', a minimum of 2 exponent digits, and a mandatory sign on the exponent, to keep the formatting as close as possible to previous output. - The output of exact midpoint values is disabled as noted above. - The integer fast-path code is changed somewhat (since we have fixed-point output and the upstream did not). - Our project style has been largely applied to the code with the exception of C99 declaration-after-statement, which has been retained as an exception to our present policy. - Most of upstream's debugging and conditionals are removed, and we use our own configure tests to determine things like uint128 availability. Changing the float output format obviously affects a number of regression tests. This patch uses an explicit setting of extra_float_digits=0 for test output that is not expected to be exactly reproducible (e.g. due to numerical instability or differing algorithms for transcendental functions). Conversions from floats to numeric are unchanged by this patch. These may appear in index expressions and it is not yet clear whether any change should be made, so that can be left for another day. This patch assumes that the only supported floating point format is now IEEE format, and the documentation is updated to reflect that. Code by me, adapting the work of Ulf Adams and other contributors. References: https://dl.acm.org/citation.cfm?id=3192369 Reviewed-by: Tom Lane, Andres Freund, Donald Dong Discussion: https://postgr.es/m/87r2el1bx6.fsf@news-spur.riddles.org.uk
2019-02-13 16:20:33 +01:00
SET extra_float_digits = 0;
Cube extension kNN support Introduce distance operators over cubes: <#> taxicab distance <-> euclidean distance <=> chebyshev distance Also add kNN support of those distances in GiST opclass. Author: Stas Kelvich
2015-12-18 12:38:27 +01:00
SELECT *, c <-> '(100, 100),(500, 500)'::cube as dist FROM test_cube ORDER BY c <-> '(100, 100),(500, 500)'::cube LIMIT 5;
Change floating-point output format for improved performance. Previously, floating-point output was done by rounding to a specific decimal precision; by default, to 6 or 15 decimal digits (losing information) or as requested using extra_float_digits. Drivers that wanted exact float values, and applications like pg_dump that must preserve values exactly, set extra_float_digits=3 (or sometimes 2 for historical reasons, though this isn't enough for float4). Unfortunately, decimal rounded output is slow enough to become a noticable bottleneck when dealing with large result sets or COPY of large tables when many floating-point values are involved. Floating-point output can be done much faster when the output is not rounded to a specific decimal length, but rather is chosen as the shortest decimal representation that is closer to the original float value than to any other value representable in the same precision. The recently published Ryu algorithm by Ulf Adams is both relatively simple and remarkably fast. Accordingly, change float4out/float8out to output shortest decimal representations if extra_float_digits is greater than 0, and make that the new default. Applications that need rounded output can set extra_float_digits back to 0 or below, and take the resulting performance hit. We make one concession to portability for systems with buggy floating-point input: we do not output decimal values that fall exactly halfway between adjacent representable binary values (which would rely on the reader doing round-to-nearest-even correctly). This is known to be a problem at least for VS2013 on Windows. Our version of the Ryu code originates from https://github.com/ulfjack/ryu/ at commit c9c3fb1979, but with the following (significant) modifications: - Output format is changed to use fixed-point notation for small exponents, as printf would, and also to use lowercase 'e', a minimum of 2 exponent digits, and a mandatory sign on the exponent, to keep the formatting as close as possible to previous output. - The output of exact midpoint values is disabled as noted above. - The integer fast-path code is changed somewhat (since we have fixed-point output and the upstream did not). - Our project style has been largely applied to the code with the exception of C99 declaration-after-statement, which has been retained as an exception to our present policy. - Most of upstream's debugging and conditionals are removed, and we use our own configure tests to determine things like uint128 availability. Changing the float output format obviously affects a number of regression tests. This patch uses an explicit setting of extra_float_digits=0 for test output that is not expected to be exactly reproducible (e.g. due to numerical instability or differing algorithms for transcendental functions). Conversions from floats to numeric are unchanged by this patch. These may appear in index expressions and it is not yet clear whether any change should be made, so that can be left for another day. This patch assumes that the only supported floating point format is now IEEE format, and the documentation is updated to reflect that. Code by me, adapting the work of Ulf Adams and other contributors. References: https://dl.acm.org/citation.cfm?id=3192369 Reviewed-by: Tom Lane, Andres Freund, Donald Dong Discussion: https://postgr.es/m/87r2el1bx6.fsf@news-spur.riddles.org.uk
2019-02-13 16:20:33 +01:00
RESET extra_float_digits;
Cube extension kNN support Introduce distance operators over cubes: <#> taxicab distance <-> euclidean distance <=> chebyshev distance Also add kNN support of those distances in GiST opclass. Author: Stas Kelvich
2015-12-18 12:38:27 +01:00
SELECT *, c <=> '(100, 100),(500, 500)'::cube as dist FROM test_cube ORDER BY c <=> '(100, 100),(500, 500)'::cube LIMIT 5;
SELECT *, c <#> '(100, 100),(500, 500)'::cube as dist FROM test_cube ORDER BY c <#> '(100, 100),(500, 500)'::cube LIMIT 5;
Fix behavior of ~> (cube, int) operator ~> (cube, int) operator was especially designed for knn-gist search. However, it appears that knn-gist search can't work correctly with current behavior of this operator when dataset contains cubes of variable dimensionality. In this case, the same value of second operator argument can point to different dimension depending on dimensionality of particular cube. Such behavior is incompatible with gist indexing of cubes, and knn-gist doesn't work correctly for it. This patch changes behavior of ~> (cube, int) operator by introducing dimension numbering where value of second argument unambiguously identifies number of dimension. With new behavior, this operator can be correctly supported by knn-gist. Relevant changes to cube operator class are also included. Backpatch to v9.6 where operator was introduced. Since behavior of ~> (cube, int) operator is changed, depending entities must be refreshed after upgrade. Such as, expression indexes using this operator must be reindexed, materialized views must be rebuilt, stored procedures and client code must be revised to correctly use new behavior. That should be mentioned in release notes. Noticed by: Tomas Vondra Author: Alexander Korotkov Reviewed by: Tomas Vondra, Andrey Borodin Discussion: https://www.postgresql.org/message-id/flat/a9657f6a-b497-36ff-e56-482a2c7e3292@2ndquadrant.com
2018-01-11 12:41:14 +01:00
-- Test sorting by coordinates
SELECT c~>1, c FROM test_cube ORDER BY c~>1 LIMIT 15; -- ascending by left bound
SELECT c~>2, c FROM test_cube ORDER BY c~>2 LIMIT 15; -- ascending by right bound
SELECT c~>3, c FROM test_cube ORDER BY c~>3 LIMIT 15; -- ascending by lower bound
SELECT c~>4, c FROM test_cube ORDER BY c~>4 LIMIT 15; -- ascending by upper bound
llow negative coordinate for ~> (cube, int) operator ~> (cube, int) operator was especially designed for knn-gist search. However, knn-gist supports only ascending ordering of results. Nevertheless it would be useful to support descending ordering by ~> (cube, int) operator. We provide workaround for that: negative coordinate give us inversed value of corresponding cube bound. Therefore, knn search using negative coordinate gives us an effect of descending ordering by cube bound. Author: Alexander Korotkov Reviewed by: Tomas Vondra, Andrey Borodin Discussion: https://www.postgresql.org/message-id/flat/a9657f6a-b497-36ff-e56-482a2c7e3292@2ndquadrant.com
2018-01-11 12:49:36 +01:00
SELECT c~>(-1), c FROM test_cube ORDER BY c~>(-1) LIMIT 15; -- descending by left bound
SELECT c~>(-2), c FROM test_cube ORDER BY c~>(-2) LIMIT 15; -- descending by right bound
SELECT c~>(-3), c FROM test_cube ORDER BY c~>(-3) LIMIT 15; -- descending by lower bound
SELECT c~>(-4), c FROM test_cube ORDER BY c~>(-4) LIMIT 15; -- descending by upper bound
Fix behavior of ~> (cube, int) operator ~> (cube, int) operator was especially designed for knn-gist search. However, it appears that knn-gist search can't work correctly with current behavior of this operator when dataset contains cubes of variable dimensionality. In this case, the same value of second operator argument can point to different dimension depending on dimensionality of particular cube. Such behavior is incompatible with gist indexing of cubes, and knn-gist doesn't work correctly for it. This patch changes behavior of ~> (cube, int) operator by introducing dimension numbering where value of second argument unambiguously identifies number of dimension. With new behavior, this operator can be correctly supported by knn-gist. Relevant changes to cube operator class are also included. Backpatch to v9.6 where operator was introduced. Since behavior of ~> (cube, int) operator is changed, depending entities must be refreshed after upgrade. Such as, expression indexes using this operator must be reindexed, materialized views must be rebuilt, stored procedures and client code must be revised to correctly use new behavior. That should be mentioned in release notes. Noticed by: Tomas Vondra Author: Alexander Korotkov Reviewed by: Tomas Vondra, Andrey Borodin Discussion: https://www.postgresql.org/message-id/flat/a9657f6a-b497-36ff-e56-482a2c7e3292@2ndquadrant.com
2018-01-11 12:41:14 +01:00
-- Same queries with sequential scan (should give the same results as above)
RESET enable_seqscan;
SET enable_indexscan = OFF;
Change floating-point output format for improved performance. Previously, floating-point output was done by rounding to a specific decimal precision; by default, to 6 or 15 decimal digits (losing information) or as requested using extra_float_digits. Drivers that wanted exact float values, and applications like pg_dump that must preserve values exactly, set extra_float_digits=3 (or sometimes 2 for historical reasons, though this isn't enough for float4). Unfortunately, decimal rounded output is slow enough to become a noticable bottleneck when dealing with large result sets or COPY of large tables when many floating-point values are involved. Floating-point output can be done much faster when the output is not rounded to a specific decimal length, but rather is chosen as the shortest decimal representation that is closer to the original float value than to any other value representable in the same precision. The recently published Ryu algorithm by Ulf Adams is both relatively simple and remarkably fast. Accordingly, change float4out/float8out to output shortest decimal representations if extra_float_digits is greater than 0, and make that the new default. Applications that need rounded output can set extra_float_digits back to 0 or below, and take the resulting performance hit. We make one concession to portability for systems with buggy floating-point input: we do not output decimal values that fall exactly halfway between adjacent representable binary values (which would rely on the reader doing round-to-nearest-even correctly). This is known to be a problem at least for VS2013 on Windows. Our version of the Ryu code originates from https://github.com/ulfjack/ryu/ at commit c9c3fb1979, but with the following (significant) modifications: - Output format is changed to use fixed-point notation for small exponents, as printf would, and also to use lowercase 'e', a minimum of 2 exponent digits, and a mandatory sign on the exponent, to keep the formatting as close as possible to previous output. - The output of exact midpoint values is disabled as noted above. - The integer fast-path code is changed somewhat (since we have fixed-point output and the upstream did not). - Our project style has been largely applied to the code with the exception of C99 declaration-after-statement, which has been retained as an exception to our present policy. - Most of upstream's debugging and conditionals are removed, and we use our own configure tests to determine things like uint128 availability. Changing the float output format obviously affects a number of regression tests. This patch uses an explicit setting of extra_float_digits=0 for test output that is not expected to be exactly reproducible (e.g. due to numerical instability or differing algorithms for transcendental functions). Conversions from floats to numeric are unchanged by this patch. These may appear in index expressions and it is not yet clear whether any change should be made, so that can be left for another day. This patch assumes that the only supported floating point format is now IEEE format, and the documentation is updated to reflect that. Code by me, adapting the work of Ulf Adams and other contributors. References: https://dl.acm.org/citation.cfm?id=3192369 Reviewed-by: Tom Lane, Andres Freund, Donald Dong Discussion: https://postgr.es/m/87r2el1bx6.fsf@news-spur.riddles.org.uk
2019-02-13 16:20:33 +01:00
SET extra_float_digits = 0;
Fix behavior of ~> (cube, int) operator ~> (cube, int) operator was especially designed for knn-gist search. However, it appears that knn-gist search can't work correctly with current behavior of this operator when dataset contains cubes of variable dimensionality. In this case, the same value of second operator argument can point to different dimension depending on dimensionality of particular cube. Such behavior is incompatible with gist indexing of cubes, and knn-gist doesn't work correctly for it. This patch changes behavior of ~> (cube, int) operator by introducing dimension numbering where value of second argument unambiguously identifies number of dimension. With new behavior, this operator can be correctly supported by knn-gist. Relevant changes to cube operator class are also included. Backpatch to v9.6 where operator was introduced. Since behavior of ~> (cube, int) operator is changed, depending entities must be refreshed after upgrade. Such as, expression indexes using this operator must be reindexed, materialized views must be rebuilt, stored procedures and client code must be revised to correctly use new behavior. That should be mentioned in release notes. Noticed by: Tomas Vondra Author: Alexander Korotkov Reviewed by: Tomas Vondra, Andrey Borodin Discussion: https://www.postgresql.org/message-id/flat/a9657f6a-b497-36ff-e56-482a2c7e3292@2ndquadrant.com
2018-01-11 12:41:14 +01:00
SELECT *, c <-> '(100, 100),(500, 500)'::cube as dist FROM test_cube ORDER BY c <-> '(100, 100),(500, 500)'::cube LIMIT 5;
Change floating-point output format for improved performance. Previously, floating-point output was done by rounding to a specific decimal precision; by default, to 6 or 15 decimal digits (losing information) or as requested using extra_float_digits. Drivers that wanted exact float values, and applications like pg_dump that must preserve values exactly, set extra_float_digits=3 (or sometimes 2 for historical reasons, though this isn't enough for float4). Unfortunately, decimal rounded output is slow enough to become a noticable bottleneck when dealing with large result sets or COPY of large tables when many floating-point values are involved. Floating-point output can be done much faster when the output is not rounded to a specific decimal length, but rather is chosen as the shortest decimal representation that is closer to the original float value than to any other value representable in the same precision. The recently published Ryu algorithm by Ulf Adams is both relatively simple and remarkably fast. Accordingly, change float4out/float8out to output shortest decimal representations if extra_float_digits is greater than 0, and make that the new default. Applications that need rounded output can set extra_float_digits back to 0 or below, and take the resulting performance hit. We make one concession to portability for systems with buggy floating-point input: we do not output decimal values that fall exactly halfway between adjacent representable binary values (which would rely on the reader doing round-to-nearest-even correctly). This is known to be a problem at least for VS2013 on Windows. Our version of the Ryu code originates from https://github.com/ulfjack/ryu/ at commit c9c3fb1979, but with the following (significant) modifications: - Output format is changed to use fixed-point notation for small exponents, as printf would, and also to use lowercase 'e', a minimum of 2 exponent digits, and a mandatory sign on the exponent, to keep the formatting as close as possible to previous output. - The output of exact midpoint values is disabled as noted above. - The integer fast-path code is changed somewhat (since we have fixed-point output and the upstream did not). - Our project style has been largely applied to the code with the exception of C99 declaration-after-statement, which has been retained as an exception to our present policy. - Most of upstream's debugging and conditionals are removed, and we use our own configure tests to determine things like uint128 availability. Changing the float output format obviously affects a number of regression tests. This patch uses an explicit setting of extra_float_digits=0 for test output that is not expected to be exactly reproducible (e.g. due to numerical instability or differing algorithms for transcendental functions). Conversions from floats to numeric are unchanged by this patch. These may appear in index expressions and it is not yet clear whether any change should be made, so that can be left for another day. This patch assumes that the only supported floating point format is now IEEE format, and the documentation is updated to reflect that. Code by me, adapting the work of Ulf Adams and other contributors. References: https://dl.acm.org/citation.cfm?id=3192369 Reviewed-by: Tom Lane, Andres Freund, Donald Dong Discussion: https://postgr.es/m/87r2el1bx6.fsf@news-spur.riddles.org.uk
2019-02-13 16:20:33 +01:00
RESET extra_float_digits;
Fix behavior of ~> (cube, int) operator ~> (cube, int) operator was especially designed for knn-gist search. However, it appears that knn-gist search can't work correctly with current behavior of this operator when dataset contains cubes of variable dimensionality. In this case, the same value of second operator argument can point to different dimension depending on dimensionality of particular cube. Such behavior is incompatible with gist indexing of cubes, and knn-gist doesn't work correctly for it. This patch changes behavior of ~> (cube, int) operator by introducing dimension numbering where value of second argument unambiguously identifies number of dimension. With new behavior, this operator can be correctly supported by knn-gist. Relevant changes to cube operator class are also included. Backpatch to v9.6 where operator was introduced. Since behavior of ~> (cube, int) operator is changed, depending entities must be refreshed after upgrade. Such as, expression indexes using this operator must be reindexed, materialized views must be rebuilt, stored procedures and client code must be revised to correctly use new behavior. That should be mentioned in release notes. Noticed by: Tomas Vondra Author: Alexander Korotkov Reviewed by: Tomas Vondra, Andrey Borodin Discussion: https://www.postgresql.org/message-id/flat/a9657f6a-b497-36ff-e56-482a2c7e3292@2ndquadrant.com
2018-01-11 12:41:14 +01:00
SELECT *, c <=> '(100, 100),(500, 500)'::cube as dist FROM test_cube ORDER BY c <=> '(100, 100),(500, 500)'::cube LIMIT 5;
SELECT *, c <#> '(100, 100),(500, 500)'::cube as dist FROM test_cube ORDER BY c <#> '(100, 100),(500, 500)'::cube LIMIT 5;
SELECT c~>1, c FROM test_cube ORDER BY c~>1 LIMIT 15; -- ascending by left bound
SELECT c~>2, c FROM test_cube ORDER BY c~>2 LIMIT 15; -- ascending by right bound
SELECT c~>3, c FROM test_cube ORDER BY c~>3 LIMIT 15; -- ascending by lower bound
SELECT c~>4, c FROM test_cube ORDER BY c~>4 LIMIT 15; -- ascending by upper bound
llow negative coordinate for ~> (cube, int) operator ~> (cube, int) operator was especially designed for knn-gist search. However, knn-gist supports only ascending ordering of results. Nevertheless it would be useful to support descending ordering by ~> (cube, int) operator. We provide workaround for that: negative coordinate give us inversed value of corresponding cube bound. Therefore, knn search using negative coordinate gives us an effect of descending ordering by cube bound. Author: Alexander Korotkov Reviewed by: Tomas Vondra, Andrey Borodin Discussion: https://www.postgresql.org/message-id/flat/a9657f6a-b497-36ff-e56-482a2c7e3292@2ndquadrant.com
2018-01-11 12:49:36 +01:00
SELECT c~>(-1), c FROM test_cube ORDER BY c~>(-1) LIMIT 15; -- descending by left bound
SELECT c~>(-2), c FROM test_cube ORDER BY c~>(-2) LIMIT 15; -- descending by right bound
SELECT c~>(-3), c FROM test_cube ORDER BY c~>(-3) LIMIT 15; -- descending by lower bound
SELECT c~>(-4), c FROM test_cube ORDER BY c~>(-4) LIMIT 15; -- descending by upper bound
Fix behavior of ~> (cube, int) operator ~> (cube, int) operator was especially designed for knn-gist search. However, it appears that knn-gist search can't work correctly with current behavior of this operator when dataset contains cubes of variable dimensionality. In this case, the same value of second operator argument can point to different dimension depending on dimensionality of particular cube. Such behavior is incompatible with gist indexing of cubes, and knn-gist doesn't work correctly for it. This patch changes behavior of ~> (cube, int) operator by introducing dimension numbering where value of second argument unambiguously identifies number of dimension. With new behavior, this operator can be correctly supported by knn-gist. Relevant changes to cube operator class are also included. Backpatch to v9.6 where operator was introduced. Since behavior of ~> (cube, int) operator is changed, depending entities must be refreshed after upgrade. Such as, expression indexes using this operator must be reindexed, materialized views must be rebuilt, stored procedures and client code must be revised to correctly use new behavior. That should be mentioned in release notes. Noticed by: Tomas Vondra Author: Alexander Korotkov Reviewed by: Tomas Vondra, Andrey Borodin Discussion: https://www.postgresql.org/message-id/flat/a9657f6a-b497-36ff-e56-482a2c7e3292@2ndquadrant.com
2018-01-11 12:41:14 +01:00
RESET enable_indexscan;