Take the 2-minute tour ×
Database Administrators Stack Exchange is a question and answer site for database professionals who wish to improve their database skills and learn from others in the community. It's 100% free, no registration required.

I have a large table of objects (15M+ row) in PostgreSQL 9.0.8, for which I want to query for outdated field.

I want to divide the query by millions, for scalability & concurrency purposes, and I want to fetch all data with the updated_at field with a date of a few days ago.

I have tried many indexes, and queries, on a million ids, and I can't seem to get performance under 100 seconds with Heroku's Ronin hardware.

I am looking for suggestions I haven't tried to make this as efficient as possible.

TRY #1

 EXPLAIN ANALYZE SELECT count(*) FROM objects
 WHERE (date(updated_at)) < (date(now())-7) AND id >= 5000001 AND id < 6000001;
 INDEX USED: (date(updated_at),id)
 268578.934 ms

TRY #2

 EXPLAIN ANALYZE SELECT count(*) FROM objects
 WHERE ((date(now()) - (date(updated_at)) > 7)) AND id >= 5000001 AND id < 6000001;
 INDEX USED: primary key
 335555.144 ms

TRY #3

 EXPLAIN ANALYZE SELECT count(*) FROM objects
 WHERE (date(updated_at)) < (date(now())-7) AND id/1000000 = 5;
 INDEX USED: (date(updated_at),(id/1000000))
 243427.042 ms

TRY #4

 EXPLAIN ANALYZE SELECT count(*) FROM objects
 WHERE (date(updated_at)) < (date(now())-7) AND id/1000000 = 5 AND updated_at IS NOT NULL;
 INDEX USED: (date(updated_at),(id/1000000)) WHERE updated_at IS NOT NULL 
 706714.812 ms

TRY #5 (for a single month of outdated data)

 EXPLAIN ANALYZE SELECT count(*) FROM objects
 WHERE (EXTRACT(MONTH from date(updated_at)) = 8) AND id/1000000 = 5;
 INDEX USED: (EXTRACT(MONTH from date(updated_at)),(id/1000000))
 107241.472 ms

TRY #6

 EXPLAIN ANALYZE SELECT count(*) FROM objects
 WHERE (date(updated_at)) < (date(now())-7) AND id/1000000 = 5;
 INDEX USED: ( (id/1000000 ) ASC ,updated_at DESC NULLS LAST)
 106842.395 ms

TRY #7 (see: http://explain.depesz.com/s/DQP)

 EXPLAIN ANALYZE SELECT count(*) FROM objects
 WHERE id/1000000 = 5 and (date(updated_at)) < (date(now())-7);
 INDEX USED: ( (id/1000000 ) ASC ,date(updated_at) DESC NULLS LAST);
 100732.049 ms
 Second try: 87280.728 ms 

TRY #8

 EXPLAIN ANALYZE SELECT count(*) FROM objects
 WHERE (date(updated_at)) < (date(now())-7) AND id/1000000 = 5 AND updated_at IS NOT NULL;
 INDEX USED:  ( (id/1000000 ) ASC ,date(updated_at) ASC NULLS LAST);
 129133.022 ms

TRY #9 (partial index as per Erwin's suggestion, see: http://explain.depesz.com/s/p9A)

 EXPLAIN ANALYZE SELECT count(*) FROM objects
 WHERE id BETWEEN 5000000 AND 5999999 AND (date(updated_at)) < '2012-10-23'::date;
 INDEX USED: (date(updated_at) DESC NULLS LAST)
 WHERE id BETWEEN 5000000 AND 6000000 AND date(updated_at) < '2012-10-23'::date;
 73861.047 ms

TRY #10 (CLUSTER, as per Erwin's suggestion).

 CREATE INDEX ix_8 on objects ( (id/1000000 ) ASC ,date(updated_at) DESC NULLS LAST);
 CLUSTER entities USING ix_8;
 EXPLAIN ANALYZE SELECT count(*) FROM objects
 WHERE id/1000000 = 5 and (date(updated_at)) < (date(now())-7) ;
 4745.595 ms

 EXPLAIN ANALYZE SELECT count(*) FROM objects
 WHERE id/1000000 = 10 and (date(updated_at)) < (date(now())-7) ;
 17573.639 ms

==> This solution seems to be the winning one. I'll have to test thoroughly to verify counterimpacts everywhere in my application.

DB settings:

select name,min_val, max_val, boot_val from pg_settings;

             name               |  min_val  |   max_val    |     boot_val      
--------------------------------+-----------+--------------+-------------------
allow_system_table_mods         |           |              | off
application_name                |           |              | 
archive_command                 |           |              | 
archive_mode                    |           |              | off
archive_timeout                 | 0         | 2147483647   | 0
array_nulls                     |           |              | on
authentication_timeout          | 1         | 600          | 60
autovacuum                      |           |              | on
autovacuum_analyze_scale_factor | 0         | 100          | 0.1
autovacuum_analyze_threshold    | 0         | 2147483647   | 50
autovacuum_freeze_max_age       | 100000000 | 2000000000   | 200000000
autovacuum_max_workers          | 1         | 536870911    | 3
autovacuum_naptime              | 1         | 2147483      | 60
autovacuum_vacuum_cost_delay    | -1        | 100          | 20
autovacuum_vacuum_cost_limit    | -1        | 10000        | -1
autovacuum_vacuum_scale_factor  | 0         | 100          | 0.2
autovacuum_vacuum_threshold     | 0         | 2147483647   | 50
backslash_quote                 |           |              | safe_encoding
bgwriter_delay                  | 10        | 10000        | 200
bgwriter_lru_maxpages           | 0         | 1000         | 100
bgwriter_lru_multiplier         | 0         | 10           | 2
block_size                      | 8192      | 8192         | 8192
bonjour                         |           |              | off
bonjour_name                    |           |              | 
bytea_output                    |           |              | hex
check_function_bodies           |           |              | on
checkpoint_completion_target    | 0         | 1            | 0.5
checkpoint_segments             | 1         | 2147483647   | 3
checkpoint_timeout              | 30        | 3600         | 300
checkpoint_warning              | 0         | 2147483647   | 30
client_encoding                 |           |              | SQL_ASCII
client_min_messages             |           |              | notice
commit_delay                    | 0         | 100000       | 0
commit_siblings                 | 1         | 1000         | 5
constraint_exclusion            |           |              | partition
cpu_index_tuple_cost            | 0         | 1.79769e+308 | 0.005
cpu_operator_cost               | 0         | 1.79769e+308 | 0.0025
cpu_tuple_cost                  | 0         | 1.79769e+308 | 0.01
cursor_tuple_fraction           | 0         | 1            | 0.1
custom_variable_classes         |           |              | 
DateStyle                       |           |              | ISO, MDY
db_user_namespace               |           |              | off
deadlock_timeout                | 1         | 2147483      | 1000
debug_assertions                |           |              | off
debug_pretty_print              |           |              | on
debug_print_parse               |           |              | off
debug_print_plan                |           |              | off
debug_print_rewritten           |           |              | off
default_statistics_target       | 1         | 10000        | 100
default_tablespace              |           |              | 
default_text_search_config      |           |              | pg_catalog.simple
default_transaction_isolation   |           |              | read committed
default_transaction_read_only   |           |              | off
default_with_oids               |           |              | off
effective_cache_size            | 1         | 2147483647   | 16384
effective_io_concurrency        | 0         | 1000         | 1
enable_bitmapscan               |           |              | on
enable_hashagg                  |           |              | on
enable_hashjoin                 |           |              | on
enable_indexscan                |           |              | on
enable_material                 |           |              | on
enable_mergejoin                |           |              | on
enable_nestloop                 |           |              | on
enable_seqscan                  |           |              | on
enable_sort                     |           |              | on
enable_tidscan                  |           |              | on
escape_string_warning           |           |              | on
extra_float_digits              | -15       | 3            | 0
from_collapse_limit             | 1         | 2147483647   | 8
fsync                           |           |              | on
full_page_writes                |           |              | on
geqo                            |           |              | on
geqo_effort                     | 1         | 10           | 5
geqo_generations                | 0         | 2147483647   | 0
geqo_pool_size                  | 0         | 2147483647   | 0
geqo_seed                       | 0         | 1            | 0
geqo_selection_bias             | 1.5       | 2            | 2
geqo_threshold                  | 2         | 2147483647   | 12
gin_fuzzy_search_limit          | 0         | 2147483647   | 0
hot_standby                     |           |              | off
ignore_system_indexes           |           |              | off
integer_datetimes               |           |              | on
IntervalStyle                   |           |              | postgres
join_collapse_limit             | 1         | 2147483647   | 8
krb_caseins_users               |           |              | off
krb_srvname                     |           |              | postgres
lc_collate                      |           |              | C
lc_ctype                        |           |              | C
lc_messages                     |           |              |
lc_monetary                     |           |              | C
lc_numeric                      |           |              | C
lc_time                         |           |              | C
listen_addresses                |           |              | localhost
lo_compat_privileges            |           |              | off
local_preload_libraries         |           |              |
log_autovacuum_min_duration     | -1        | 2147483      | -1
log_checkpoints                 |           |              | off
log_connections                 |           |              | off
log_destination                 |           |              | stderr
log_disconnections              |           |              | off
log_duration                    |           |              | off
log_error_verbosity             |           |              | default
log_executor_stats              |           |              | off
log_hostname                    |           |              | off
log_line_prefix                 |           |              |
log_lock_waits                  |           |              | off
log_min_duration_statement      | -1        | 2147483      | -1
log_min_error_statement         |           |              | error
log_min_messages                |           |              | warning
log_parser_stats                |           |              | off
log_planner_stats               |           |              | off
log_rotation_age                | 0         | 35791394     | 1440
log_rotation_size               | 0         | 2097151      | 10240
log_statement                   |           |              | none
log_statement_stats             |           |              | off
log_temp_files                  | -1        | 2147483647   | -1
log_timezone                    |           |              | UNKNOWN
log_truncate_on_rotation        |           |              | off
logging_collector               |           |              | off
maintenance_work_mem            | 1024      | 2097151      | 16384
max_connections                 | 1         | 536870911    | 100
max_files_per_process           | 25        | 2147483647   | 1000
max_function_args               | 100       | 100          | 100
max_identifier_length           | 63        | 63           | 63
max_index_keys                  | 32        | 32           | 32
max_locks_per_transaction       | 10        | 2147483647   | 64
max_prepared_transactions       | 0         | 536870911    | 0
max_stack_depth                 | 100       | 2097151      | 100
max_standby_archive_delay       | -1        | 2147483      | 30000
max_standby_streaming_delay     | -1        | 2147483      | 30000
max_wal_senders                 | 0         | 536870911    | 0
password_encryption             |           |              | on
port                            | 1         | 65535        | 5432
post_auth_delay                 | 0         | 2147483647   | 0
pre_auth_delay                  | 0         | 60           | 0
random_page_cost                | 0         | 1.79769e+308 | 4
search_path                     |           |              | "$user",public
segment_size                    | 131072    | 131072       | 131072
seq_page_cost                   | 0         | 1.79769e+308 | 1
server_encoding                 |           |              | SQL_ASCII
server_version                  |           |              | 9.0.8
server_version_num              | 90008     | 90008        | 90008
session_replication_role        |           |              | origin
shared_buffers                  | 16        | 1073741823   | 1024
silent_mode                     |           |              | off
sql_inheritance                 |           |              | on
ssl                             |           |              | off
ssl_renegotiation_limit         | 0         | 2097151      | 524288
standard_conforming_strings     |           |              | off
statement_timeout               | 0         | 2147483647   | 0
superuser_reserved_connections  | 0         | 536870911    | 3
synchronize_seqscans            |           |              | on
synchronous_commit              |           |              | on
syslog_facility                 |           |              | local0
syslog_ident                    |           |              | postgres
tcp_keepalives_count            | 0         | 2147483647   | 0
tcp_keepalives_idle             | 0         | 2147483647   | 0
tcp_keepalives_interval         | 0         | 2147483647   | 0
temp_buffers                    | 100       | 1073741823   | 1024
temp_tablespaces                |           |              |
TimeZone                        |           |              | UNKNOWN
timezone_abbreviations          |           |              | UNKNOWN
trace_notify                    |           |              | off
trace_recovery_messages         |           |              | log
trace_sort                      |           |              | off
track_activities                |           |              | on
track_activity_query_size       | 100       | 102400       | 1024
track_counts                    |           |              | on
track_functions                 |           |              | none
transaction_isolation           |           |              |
transaction_read_only           |           |              | off
transform_null_equals           |           |              | off
unix_socket_group               |           |              |
unix_socket_permissions         | 0         | 511          | 511
update_process_title            |           |              | on
vacuum_cost_delay               | 0         | 100          | 0
vacuum_cost_limit               | 1         | 10000        | 200
vacuum_cost_page_dirty          | 0         | 10000        | 20
vacuum_cost_page_hit            | 0         | 10000        | 1
vacuum_cost_page_miss           | 0         | 10000        | 10
vacuum_defer_cleanup_age        | 0         | 1000000      | 0
vacuum_freeze_min_age           | 0         | 1000000000   | 50000000
vacuum_freeze_table_age         | 0         | 2000000000   | 150000000
wal_block_size                  | 8192      | 8192         | 8192
wal_buffers                     | 4         | 2147483647   | 8
wal_keep_segments               | 0         | 2147483647   | 0
wal_level                       |           |              | minimal
wal_segment_size                | 2048      | 2048         | 2048
wal_sender_delay                | 1         | 10000        | 200
wal_sync_method                 |           |              | fdatasync
wal_writer_delay                | 1         | 10000        | 200
work_mem                        | 64        | 2097151      | 1024
xmlbinary                       |           |              | base64
xmloption                       |           |              | content
zero_damaged_pages              |           |              | off
(195 rows)
share|improve this question

migrated from stackoverflow.com Oct 31 '12 at 1:30

This question came from our site for professional and enthusiast programmers.

    
Maybe you could give some additional information: table structure, index definition, expected proportion of outdated dates, row_width, tuning parameters? The solution will probably be either tuning, or partial indices. –  wildplasser Oct 29 '12 at 22:33
    
The table structure has ~20-30 columns, a few foreign keys integers, string, text, booleans. The index definition is in the posting above next to INDEX USEd. (I posted the 8 indexes used by the queries). I have a few more indexes for quicker update and selects for my application. Finally, I am using a cloud DB, and I haven't change anything. I am wondering mostly if my index definition is as good as it can gets before getting to those optimization. Nevertheless, I'll update with the info. –  xlash Oct 29 '12 at 22:43
    
Ok. Thanks Thanks. (BTW Most gurus read SQL faster than natural language text ;-) I's also less ambiguous... BTW2: I am not a guru, but I least I have that habit. BTW3 15M rows is not huge. Only a few thousand pages, given moderate rowsize. –  wildplasser Oct 29 '12 at 22:45
1  
If I understand correctly you can add the context field to the pg_settings query to find out if you can change that particular constant: select name,context,min_val, max_val, boot_val from pg_settings;. If context=='user', you could use SET xxx = val; to set a parameter before the query. for xxx= {effective_cache_size, work_mem, random_page_cost} are the usual suspects. –  wildplasser Oct 29 '12 at 23:43
    
Thanks for showing your Pg version. While it's good that you've posted real query timings, what'd be really good would be to show the DDL (as wildplasser asked) and the actual query plan for at least some of the queries. Paste the explain analyze output to explain.depesz.com and link to it in the answer. –  Craig Ringer Oct 30 '12 at 0:06

1 Answer 1

First off, can it be? You write:

I want to fetch all data with the updated_at field with a date of a few days ago.

But your WHERE condition is:

(date(updated_at)) < (date(now())-7)

Shouldn't that be >?


Indexes

For optimal performance, you could ...

  • partition your indexes
  • exclude irrelevant rows from the indexes
  • automatically recreate indexes at off-hours with updated predicate.

Your indexes could look like:

CREATE INDEX objects_id_updated_at_idx (updated_at::date DESC NULLS LAST)
WHERE  id BETWEEN 0 AND 999999
AND    updated_at > '2012-10-01 0:0'::timestamp  -- some minimum date

CREATE INDEX objects_id_updated_at_idx (updated_at::date DESC NULLS LAST)
WHERE  id BETWEEN 1000000 AND 1999999
AND    updated_at > '2012-10-01 0:0'::timestamp  -- some minimum date

...

The second condition excludes irrelevant rows from the index right away, which should make it smaller and faster - depending on your actual data distribution. In accordance with my preliminary comment, I am assuming you want newer rows.

The condition also automatically excludes NULL values in updated_at - which you seem to allow in the table and obviously want to exclude in the query. The usefulness of the index deteriorates over time. The query always retrieves the latest entries. Recreate the index with an updated WHERE clause periodically. This requires an exclusive lock on the table, so do it at off hours. There is also CREATE INDEX CONCURRENTLY to minimize the duration of the lock:

CREATE INDEX CONCURRENTLY objects_id_up_201211_idx; -- create new idx
DROP INDEX  objects_id_up_201210_idx;  -- then drop old

Related answer on SO:

To further optimize, you could use CLUSTER like we mentioned in the comments. But you need a full index for that. Doesn't work with a partial index. You would create temporarily:

CREATE INDEX objects_full_idx (id/1000000, updated_at::date DESC NULLS LAST);

This form of the full index matches the sort order of above partial indexes.

CLUSTER objects USING objects_full_idx;
ANALYZE objects;

This will take a while, since the table is rewritten physically. It is also effectively a VACUUM FULL. It needs an exclusive write lock on the table, so do it at off-hours - provided you can afford that at all. Again, there is a less invasive alternative: pg_repack

You can then drop the index again. It's a one-time effect. I would at least try this once to see how much your queries benefit from it. The effect deteriorates with subsequent write operations. You could repeat this procedure at off hours if you see a considerable effect.

If your table receives a lot of write operations, you have to weigh cost and benefit for this step. For many UPDATEs consider setting a FILLFACTOR lower than 100. Do that before you CLUSTER.

Query

SELECT count(*)
FROM   objects
WHERE  id BETWEEN 0 AND 999999  -- match conditions of partial index!
AND    updated_at > '2012-10-01 0:0'::timestamp
AND    updated_at::date > (now()::date - 7)

More

This related answer presents a more advanced technique for index partitioning:

Among other things it provides example code for automatic index (re-)creation.

PostgreSQL 9.2+ has several new features for you. Index-only scans alone would make it worth your while.

Make sure that autovacuum is running properly. The huge gain by CLUSTER you have reported may be due in part to the implicit VACUUM FULL that you get from CLUSTER. Maybe this is set up by Heroku automatically, not sure.
The settings in your question look good. So that's probably not an issue here and CLUSTER really was that effective.

share|improve this answer
    
So far so good ! With Partial index : 73861.047ms I'll continue the CLUSTER test later on. Thanks –  xlash Oct 30 '12 at 16:53
    
And for now, Heroku doesn't support 9.2 yet. I can upgrade to 9.1 only. –  xlash Oct 30 '12 at 18:06
    
@xlash: Note that I had a missing 0 in the full index. –  Erwin Brandstetter Oct 30 '12 at 21:42
    
With cluster, performance is 10x better. See results above (update of main question). I'll have extended tests to perform to verify other impacts –  xlash Oct 31 '12 at 2:40
    
@xlash: You can still use the partial indexes (and matching queries) after CLUSTER - which I designed to agree with the partial indexes). Should be faster than using the full index, which is only needed for CLUSTER. Then all of your queries should be equally fast (like the faster one in your try #10), given they return an equal number of rows. Also, I added a hint concerning autovacuum. –  Erwin Brandstetter Oct 31 '12 at 19:15

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.