mirror of
https://github.com/timescale/timescaledb.git
synced 2025-05-20 12:47:40 +08:00
e30699101b
The decision to add a minmax sparse index is made every time when the compressed chunk is created (full decompression followed by compression) based on the currently present indexes on the hypertable. No new chunk compression settings are added. No action is required on upgrade, but the feature is not enabled on existing chunks. The minmax index will be added when the chunk is fully decompressed and compressed. No action is required on downgrade, we ignore the unknown metadata columns. They will be removed when the chunk is fully decompressed and compressed. The potential drawback of this feature is increasing the storage requirements for the compressed chunk table, but it is normally only a few percent of the total compressed data size. It can be disabled with the GUC `timescaledb.auto_sparse_indexes`. Here's a small example of this feature in action: https://gist.github.com/akuzm/84d4b3b609e3581768173bd21001dfbf Note that the number of hit buffers is reduced almost 4x.
67 lines
2.3 KiB
SQL
67 lines
2.3 KiB
SQL
-- This file and its contents are licensed under the Timescale License.
|
|
-- Please see the included NOTICE for copyright information and
|
|
-- LICENSE-TIMESCALE for a copy of the license.
|
|
|
|
create table sparse(ts int, value float);
|
|
select create_hypertable('sparse', 'ts');
|
|
insert into sparse select x, x from generate_series(1, 10000) x;
|
|
alter table sparse set (timescaledb.compress);
|
|
|
|
-- When the chunks are compressed, minmax metadata are created for columns that
|
|
-- have btree indexes.
|
|
create index ii on sparse(value);
|
|
select count(compress_chunk(x)) from show_chunks('sparse') x;
|
|
explain select * from sparse where value = 1;
|
|
|
|
|
|
-- Should be disabled with the GUC
|
|
set timescaledb.auto_sparse_indexes to off;
|
|
select count(compress_chunk(decompress_chunk(x))) from show_chunks('sparse') x;
|
|
explain select * from sparse where value = 1;
|
|
reset timescaledb.auto_sparse_indexes;
|
|
select count(compress_chunk(decompress_chunk(x))) from show_chunks('sparse') x;
|
|
explain select * from sparse where value = 1;
|
|
|
|
|
|
-- Should survive renames.
|
|
alter table sparse rename column value to wert;
|
|
explain select * from sparse where wert = 1;
|
|
alter table sparse rename column wert to value;
|
|
explain select * from sparse where value = 1;
|
|
|
|
|
|
-- Not for expression indexes.
|
|
drop index ii;
|
|
create index ii on sparse((value + 1));
|
|
select count(compress_chunk(decompress_chunk(x))) from show_chunks('sparse') x;
|
|
explain select * from sparse where value = 1;
|
|
|
|
|
|
-- Not for other index types.
|
|
drop index ii;
|
|
create index ii on sparse using hash(value);
|
|
select count(compress_chunk(decompress_chunk(x))) from show_chunks('sparse') x;
|
|
explain select * from sparse where value = 1;
|
|
|
|
|
|
-- When the chunk is recompressed without index, no sparse index is created.
|
|
drop index ii;
|
|
select count(compress_chunk(decompress_chunk(x))) from show_chunks('sparse') x;
|
|
explain select * from sparse where value = 1;
|
|
|
|
|
|
-- Long column names.
|
|
select count(decompress_chunk(x)) from show_chunks('sparse') x;
|
|
|
|
\set ECHO none
|
|
select format('alter table sparse add column %1$s int; create index on sparse(%1$s);',
|
|
substr('Abcdef012345678_Bbcdef012345678_Cbcdef012345678_Dbcdef012345678_', 1, x))
|
|
from generate_series(1, 63) x
|
|
\gexec
|
|
\set ECHO queries
|
|
|
|
select count(compress_chunk(x)) from show_chunks('sparse') x;
|
|
|
|
explain select * from sparse where Abcdef012345678_Bbcdef012345678_Cbcdef012345678_Dbcdef0 = 1;
|
|
|