|
| 1 | +-- Use it to generate a database migration (e.g. RoR's db:migrate or Sqitch) |
| 2 | +-- to drop unused and redundant indexes. |
| 3 | + |
| 4 | +-- This query generates a set of `DROP INDEX` statements, that |
| 5 | +-- can be used in your migration script. Also, it generates |
| 6 | +-- `CREATE INDEX`, put them to revert/rollback migration script. |
| 7 | + |
| 8 | +-- It is also a good idea to manually double check all indexes being dropped. |
| 9 | +-- WARNING here: when you are dropping an index B which is redundant to some index A, |
| 10 | +-- check that you don't drop the A itself at the same time (it can be in "unused"). |
| 11 | +-- So if B is "redundant" to A and A is "unused", the script will suggest |
| 12 | +-- dropping both. If so, it is probably better to drop B and leave A. |
| 13 | +-- -- in this case there is a chance that A will be used. If it will still be unused, |
| 14 | +-- you will drop it during the next cleanup routine procedure. |
| 15 | + |
| 16 | +-- This query doesn't need any additional extensions to be installed |
| 17 | +-- (except plpgsql), and doesn't create anything (like views or smth) |
| 18 | +-- -- so feel free to use it in your clouds (Heroku, AWS RDS, etc) |
| 19 | + |
| 20 | +-- It also does't do anything except reading system catalogs and |
| 21 | +-- printing NOTICEs, so you can easily run it on your |
| 22 | +-- production *master* database. |
| 23 | +-- (Keep in mind, that on replicas, the whole picture of index usage |
| 24 | +-- is usually very different from master). |
| 25 | + |
| 26 | +-- TODO: take into account type of index and opclass |
| 27 | +-- TODO: schemas |
| 28 | + |
| 29 | +with unused as ( |
| 30 | + select |
| 31 | + format('unused (idx_scan: %s)', pg_stat_user_indexes.idx_scan)::text as reason, |
| 32 | + pg_stat_user_indexes.relname as tablename, |
| 33 | + pg_stat_user_indexes.schemaname || '.' || indexrelname::text as indexname, |
| 34 | + pg_stat_user_indexes.idx_scan, |
| 35 | + (coalesce(n_tup_ins, 0) + coalesce(n_tup_upd, 0) - coalesce(n_tup_hot_upd, 0) + coalesce(n_tup_del, 0)) as write_activity, |
| 36 | + pg_stat_user_tables.seq_scan, |
| 37 | + pg_stat_user_tables.n_live_tup, |
| 38 | + pg_get_indexdef(pg_index.indexrelid) as indexdef, |
| 39 | + pg_size_pretty(pg_relation_size(pg_index.indexrelid::regclass)) as size, |
| 40 | + pg_index.indexrelid |
| 41 | + from pg_stat_user_indexes |
| 42 | + join pg_stat_user_tables |
| 43 | + on pg_stat_user_indexes.relid = pg_stat_user_tables.relid |
| 44 | + join pg_index |
| 45 | + ON pg_index.indexrelid = pg_stat_user_indexes.indexrelid |
| 46 | + where |
| 47 | + pg_stat_user_indexes.idx_scan = 0 /* < 10 or smth */ |
| 48 | + and pg_index.indisunique is false |
| 49 | + and pg_stat_user_indexes.idx_scan::float/(coalesce(n_tup_ins,0)+coalesce(n_tup_upd,0)-coalesce(n_tup_hot_upd,0)+coalesce(n_tup_del,0)+1)::float<0.01 |
| 50 | + and (coalesce(n_tup_ins,0)+coalesce(n_tup_upd,0)-coalesce(n_tup_hot_upd,0)+coalesce(n_tup_del,0))>10000 |
| 51 | +), index_data as ( |
| 52 | + select *, string_to_array(indkey::text,' ') as key_array,array_length(string_to_array(indkey::text,' '),1) as nkeys |
| 53 | + from pg_index |
| 54 | +), redundant as ( |
| 55 | + select |
| 56 | + format('redundant to index: %I', i1.indexrelid::regclass)::text as reason, |
| 57 | + i2.indrelid::regclass::text as tablename, |
| 58 | + i2.indexrelid::regclass::text as indexname, |
| 59 | + pg_get_indexdef(i1.indexrelid) main_indexdef, |
| 60 | + pg_get_indexdef(i2.indexrelid) indexdef, |
| 61 | + pg_size_pretty(pg_relation_size(i2.indexrelid)) size, |
| 62 | + i2.indexrelid |
| 63 | + from |
| 64 | + index_data as i1 |
| 65 | + join index_data as i2 on i1.indrelid = i2.indrelid and i1.indexrelid <> i2.indexrelid |
| 66 | + where |
| 67 | + (regexp_replace(i1.indpred, 'location \d+', 'location', 'g') IS NOT DISTINCT FROM regexp_replace(i2.indpred, 'location \d+', 'location', 'g')) |
| 68 | + and (regexp_replace(i1.indexprs, 'location \d+', 'location', 'g') IS NOT DISTINCT FROM regexp_replace(i2.indexprs, 'location \d+', 'location', 'g')) |
| 69 | + and ((i1.nkeys > i2.nkeys and not i2.indisunique) OR (i1.nkeys=i2.nkeys and ((i1.indisunique and i2.indisunique and (i1.indexrelid>i2.indexrelid)) or (not i1.indisunique and not i2.indisunique and (i1.indexrelid>i2.indexrelid)) or (i1.indisunique and not i2.indisunique)))) |
| 70 | + and i1.key_array[1:i2.nkeys]=i2.key_array |
| 71 | +), together as ( |
| 72 | + select reason, tablename, indexname, size, indexdef, null as main_indexdef, indexrelid |
| 73 | + from unused |
| 74 | + union all |
| 75 | + select reason, tablename, indexname, size, indexdef, main_indexdef, indexrelid |
| 76 | + from redundant |
| 77 | + order by tablename asc, indexname |
| 78 | +), droplines as ( |
| 79 | + select format('DROP INDEX %s; -- %s, %s, table %s', max(indexname), max(size), string_agg(reason, ', '), tablename) as line |
| 80 | + from together t1 |
| 81 | + group by tablename, indexrelid |
| 82 | + order by tablename, indexrelid |
| 83 | +), createlines as ( |
| 84 | + select format('%s; -- table %s', max(indexdef), tablename) as line |
| 85 | + from together t2 |
| 86 | + group by tablename, indexrelid |
| 87 | + order by tablename, indexrelid |
| 88 | +) |
| 89 | +select '-- Do migration: --' as out |
| 90 | +union all |
| 91 | +select * from droplines |
| 92 | +union all |
| 93 | +select '' |
| 94 | +union all |
| 95 | +select '-- Revert migration: --' |
| 96 | +union all |
| 97 | +select * from createlines; |
| 98 | + |
0 commit comments