citus
citus copied to clipboard
Distributed PostgreSQL as an extension
Steps to reproduce: ``` SQL CREATE TABLE dist_table (id int); SELECT create_distributed_table('dist_table', 'id'); SELECT count(*) FROM dist_table WHERE EXISTS (SELECT count(*) FROM pg_catalog.pg_stats_ext_exprs); NOTICE: executing the command locally: SELECT count(*)...
Doesn't reproduce on 12.1. Reproduces on 13.0/13.1/13.2 with PG17.5. Also tried 13.2 but with PG16.8 and it didn't reproduce this time, interesting .. ```c ==3228== VALGRINDERROR-BEGIN ==3228== Invalid read of...
Docker [hub](https://hub.docker.com/r/citusdata/citus/tags) misses 13.1.0
Currently, the documentation for the Citus API only includes create_distributed_table() and not create_distributed_table_concurrently[0]. The function in question is only documented in the README[1] and in a blog post[2]. We should...
Hi! I see the new release [v13.1.0](https://github.com/citusdata/citus/releases/tag/v13.1.0) on github, but I can't find anything about it on [citusdata.com](https://www.citusdata.com). The latest release mentioned at [citusdata.com/updates](https://www.citusdata.com/updates) is v13.0 (with no information about...
Currently, Citus can successfully execute queries with recurring left join with a reference to outer query via recursive planning. However, when the same subquery uses an INNER JOIN, recursive planning...
Generated by sqlsmith ```SQL select subq_1.c1 as c0, subq_1.c3 as c1, (select object_names from citus.pg_dist_object limit 1 offset 1) as c2, subq_1.c2 as c3, subq_1.c5 as c4, subq_1.c2 as c5,...
DESCRIPTION: Add `citus_stat_user_tables` UDF This UDF provides `n_live_tup`, `n_dead_tup`, `n_tup_ins` etc for each Citus table in the database, based on the definitions of these columns in the Postgres view `pg_stat_user_tables`.
Creating distributed tables have an inefficiency that causes degradation of performance when shards counts increase. The performance seems to have `O(n^2)` characteristics. Shard creation is easily assumed a non performance...
Previous approach could have caused problems with other extensions that are using queryDesc->totaltime during query execution. Use another way to count the time used before postgres' executor.