473,654 Members | 3,114 Online
Bytes | Software Development & Data Engineering Community
+ Post

Home Posts Topics Members FAQ

Optimizer choosing smaller index instead of right one

Hello,

This is probably a common problem, but I couldn't really find a direct answer in
the archives (or maybe just couldn't find one that satisfied me ;-). I created
an index specifically to help a query and the optimizer does not use it. It
prefers an older index that has one less column, but that yields much poorer
performance.

The query get the smallest date in a period, for some criteria. The one table
used in the query has 604494 rows. The index I created to try to optimize it is
"xie3cham_servi co", and has all the three columns of the table that are used in
the query. But the index being used is actually xie2cham_servic o (as can be seen
below) which has only two columns, one of them completely useless for the query!
I'm probably missing some setting, but I couldn't figure out which.

One last comment, if I test this on a server all by myself, after some test
(when, I imagine, all related disk pages are on cache) the time difference
between the two situations is much smaller (4secs against 2 secs), but, of
course, this doesn't help me much ...

Here is the associated information:

Table "public.cham_se rvico"
Column | Type | Modifiers
------------------+--------------------------------+-----------
dt_inicial | timestamp(0) without time zone | not null
cod_bxs | integer | not null
chave | integer | not null
identificacao | integer | not null
cod_cifra | character varying(25) | not null
dt_identificaca o | timestamp(0) without time zone | not null
dt_ocupacao | timestamp(0) without time zone |
temp_troca | integer |
temp_ura | integer |
temp_filapa | integer |
temp_atendpa | integer |
temp_total | integer |
dt_finalizacao | timestamp(0) without time zone |
pa | integer |
flg_tipodiscage m | character(1) |
cod_liberservic o | integer |
dt_login | timestamp(0) without time zone |
id_chave_grupo | character varying(20) |
Indexes: xpkcham_servico primary key btree (dt_inicial, cod_bxs, chave,
identificacao, cod_cifra, dt_identificaca o),
xie1cham_servic o btree (dt_finalizacao , cod_cifra),
xie2cham_servic o btree (id_chave_grupo , dt_inicial),
xie3cham_servic o btree (id_chave_grupo , cod_bxs, dt_finalizacao)
Foreign Key constraints: rchamadaxservic o FOREIGN KEY (dt_inicial, cod_bxs,
chave, identificacao) REFERENCES cham_chamada(dt _inicial, cod_bxs, chave,
identificacao) ON UPDATE NO ACTION ON DELETE NO ACTION

explain analyze SELECT dt_finalizacao FROM cham_servico WHERE id_chave_grupo =
'7458' AND cod_bxs = 1 AND dt_finalizacao > to_timestamp
('2004-04-20','YYYY-MM-DD') AND dt_finalizacao < now() ORDER BY dt_finalizacao
ASC LIMIT 1;

QUERY PLAN

--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Limit (cost=15663.77. .15663.77 rows=1 width=8) (actual time=14002.66.. 14002.66
rows=1 loops=1)
-> Sort (cost=15663.77. .15663.83 rows=26 width=8) (actual
time=14002.65.. 14002.65 rows=2 loops=1)
Sort Key: dt_finalizacao
-> Index Scan using xie2cham_servic o on cham_servico
(cost=0.00..156 63.15 rows=26 width=8) (actual time=13604.37.. 13980.16 rows=3173
loops=1)
Index Cond: (id_chave_grupo = '7458'::charact er varying)
Filter: ((cod_bxs = 1) AND ((dt_finalizaca o)::timestamp with time
zone > to_timestamp('2 004-04-20'::text, 'YYYY-MM-DD'::text)) AND
((dt_finalizaca o)::timestamp with time zone < now()))
Total runtime: 45014.06 msec
(7 rows)

Time: 45105,73 ms

If I "drop index xie2cham_servic o;", I get :

explain analyze SELECT dt_finalizacao FROM cham_servico WHERE id_chave_grupo =
'7458' AND cod_bxs = 1 AND dt_finalizacao > to_timestamp
('2004-04-20','YYYY-MM-DD') AND dt_finalizacao < now() ORDER BY dt_finalizacao
ASC LIMIT 1;

QUERY PLAN

--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Limit (cost=17600.92. .17600.92 rows=1 width=8) (actual time=8750.64..8 750.64
rows=1 loops=1)
-> Sort (cost=17600.92. .17600.99 rows=26 width=8) (actual
time=8750.63..8 750.63 rows=2 loops=1)
Sort Key: dt_finalizacao
-> Index Scan using xie3cham_servic o on cham_servico
(cost=0.00..176 00.30 rows=26 width=8) (actual time=8432.37..8 743.29 rows=3176
loops=1)
Index Cond: ((id_chave_grup o = '7458'::charact er varying) AND
(cod_bxs = 1))
Filter: (((dt_finalizac ao)::timestamp with time zone >
to_timestamp('2 004-04-20'::text, 'YYYY-MM-DD'::text)) AND
((dt_finalizaca o)::timestamp with time zone < now()))
Total runtime: 8751.50 msec
(7 rows)

Time: 8758,36 ms

Some more environment info:

Postgres 7.3.4 on Linux Redhat 9.

show all:

name | setting
--------------------------------+---------------------------------------
australian_time zones | off
authentication_ timeout | 60
autocommit | on
checkpoint_segm ents | 3
checkpoint_time out | 300
client_encoding | SQL_ASCII
client_min_mess ages | notice
commit_delay | 0
commit_siblings | 5
cpu_index_tuple _cost | 0.001
cpu_operator_co st | 0.0025
cpu_tuple_cost | 0.01
DateStyle | ISO with US (NonEuropean) conventions
db_user_namespa ce | off
deadlock_timeou t | 1000
debug_pretty_pr int | off
debug_print_par se | off
debug_print_pla n | off
debug_print_rew ritten | off
default_statist ics_target | 10
default_transac tion_isolation | read committed
dynamic_library _path | $libdir
effective_cache _size | 1000
enable_hashjoin | on
enable_indexsca n | on
enable_mergejoi n | on
enable_nestloop | on
enable_seqscan | off
enable_sort | on
enable_tidscan | on
explain_pretty_ print | on
fixbtree | on
fsync | on
geqo | on
geqo_effort | 1
geqo_generation s | 0
geqo_pool_size | 0
geqo_random_see d | -1
geqo_selection_ bias | 2
geqo_threshold | 11
hostname_lookup | off
krb_server_keyf ile | FILE:/etc/sysconfig/pgsql/krb5.keytab
lc_messages | pt_BR
lc_monetary | pt_BR
lc_numeric | pt_BR
lc_time | pt_BR
log_connections | off
log_duration | off
log_min_error_s tatement | panic
log_pid | off
log_statement | off
log_timestamp | off
max_connections | 32
max_expr_depth | 10000
max_files_per_p rocess | 1000
max_fsm_pages | 10000
max_fsm_relatio ns | 1000
max_locks_per_t ransaction | 64
password_encryp tion | on
port | 5432
pre_auth_delay | 0
random_page_cos t | 4
search_path | $user,public
server_encoding | SQL_ASCII
server_min_mess ages | notice
shared_buffers | 64
show_executor_s tats | off
show_parser_sta ts | off
show_planner_st ats | off
show_source_por t | off
show_statement_ stats | off
silent_mode | off
sort_mem | 1024
sql_inheritance | on
ssl | off
statement_timeo ut | 0
stats_block_lev el | off
stats_command_s tring | off
stats_reset_on_ server_start | on
stats_row_level | off
stats_start_col lector | on
superuser_reser ved_connections | 2
syslog | 0
syslog_facility | LOCAL0
syslog_ident | postgres
tcpip_socket | on
TimeZone | unknown
trace_notify | off
TRANSACTION ISOLATION LEVEL | READ COMMITTED
transform_null_ equals | off
unix_socket_dir ectory | unset
unix_socket_gro up | unset
unix_socket_per missions | 511
vacuum_mem | 8192
virtual_host | unset
wal_buffers | 8
wal_debug | 0
wal_sync_method | fdatasync
zero_damaged_pa ges | off

---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to ma*******@postg resql.org

Nov 23 '05 #1
0 934

This thread has been closed and replies have been disabled. Please start a new discussion.

Similar topics

2
2098
by: Carlo Paccanoni | last post by:
I tried this: use northwind go SELECT OrderDate FROM Orders WHERE OrderDate > '19950101' see the query plan? ok
3
1591
by: Philip Yale | last post by:
I'm very puzzled by the choice of NC index being made by the optimizer in this example. I don't actually think it should use an NC index at all. I have: Table: CustomerStatus_T Single data page 19 records Clustered Index on CustomerStatusID:
2
1411
by: Greg Stark | last post by:
Hm, here's a query where the optimizer is choosing the wrong plan by far. I think it boils down to it guessing wrong on how selective an rtree index is, which I guess would be hard to predict. Except if it guesses wrong by assuming it isn't selective it would be maybe 50% slower doing lots of index lookups instead of a more efficient full table scan and join. If it guesses wrong by assuming it'll be very selective as it is in this case...
4
2096
by: Fan Ruo Xin | last post by:
Hi, I have been watching MDC for a long time ... I have seen it from INDEX-SCAN to INDEX ONLY for the DISTINCT (block index columns) operator .... But the Elapsed time shows MDC still run slower. This did confuse me. Of course I expected it should run faster, because of the small block index tree, compare with the RID index tree. select distinct(int(l_shipdate)/100) from lineitem;
3
2083
by: Peter Arrenbrecht | last post by:
Hi all We ran into a very annoying optimizer problem recently. We had added a new key and self-join relation to a large production table. That key will be filled very rarely and having just being added was practically never filled in when the first user tried to delete a row from the table. Now, the optimizer tried to enforce RI on the new relation. But instead of using the index generated by the relation, it used a table scan!
1
1792
by: Mike L. Bell | last post by:
Pulling our hair out over this one. We have two databases. One is used for OLTP. The other hosts old closed transactions and is mainly used for reporting. Every month we identify rows in the primary (OLTP) database that are eligible to be moved and ship them over. Been doing this for over a year without any problems. After last weekend's move everything went sour on our history database. A set of canned queries started taking orders of...
6
1568
by: Ian Ribas | last post by:
Hello, This is probably a common problem, but I couldn't really find a direct answer in the archives (or maybe just couldn't find one that satisfied me ;-). I created an index specifically to help a query and the optimizer does not use it. It prefers an older index that has one less column, but that yields much poorer performance. The query get the smallest date in a period, for some criteria. The one table used in the query has...
5
9840
by: Venkatesh Subbaramu | last post by:
Hi Expert DBAs, I am facing an issue with a query, can anyone help me through? I have a Query: UPDATE VMUS.MSG_MSTR a SET MSG_STA_ID =5 WHERE a.PROC_ID =2 AND a.file_type_id =1 AND MSG_STA_ID IN (1, 3) AND DATA_REQ_DT IS NOT NULL We have 2 indexes for the table: Indname Column names
3
2373
by: aj | last post by:
DB2 LUW v8.2 FP 14 RHAS 2.1 Sorry if these are newbie questions. Optimizer stuff is black magic to me. For both of these, assume stats are current and an even distribution of data. ------------------------- Lets say I have a table FOO1 that has, among other columns, a column named A. There is a non-unique index on A that has medium selectivity. Lets also say I have a table FOO2 that has, among other columns, a column named B. ...
0
1661
by: P. Adhia | last post by:
Hi, Consider the following query select * from t1 where c1 = ? and c2 = ?
0
8294
by: Hystou | last post by:
Most computers default to English, but sometimes we require a different language, especially when relocating. Forgot to request a specific language before your computer shipped? No problem! You can effortlessly switch the default language on Windows 10 without reinstalling. I'll walk you through it. First, let's disable language synchronization. With a Microsoft account, language settings sync across devices. To prevent any complications,...
0
8816
Oralloy
by: Oralloy | last post by:
Hello folks, I am unable to find appropriate documentation on the type promotion of bit-fields when using the generalised comparison operator "<=>". The problem is that using the GNU compilers, it seems that the internal comparison operator "<=>" tries to promote arguments from unsigned to signed. This is as boiled down as I can make it. Here is my compilation command: g++-12 -std=c++20 -Wnarrowing bit_field.cpp Here is the code in...
0
8709
jinu1996
by: jinu1996 | last post by:
In today's digital age, having a compelling online presence is paramount for businesses aiming to thrive in a competitive landscape. At the heart of this digital strategy lies an intricately woven tapestry of website design and digital marketing. It's not merely about having a website; it's about crafting an immersive digital experience that captivates audiences and drives business growth. The Art of Business Website Design Your website is...
1
8494
by: Hystou | last post by:
Overview: Windows 11 and 10 have less user interface control over operating system update behaviour than previous versions of Windows. In Windows 11 and 10, there is no way to turn off the Windows Update option using the Control Panel or Settings app; it automatically checks for updates and installs any it finds, whether you like it or not. For most users, this new feature is actually very convenient. If you want to control the update process,...
0
8596
tracyyun
by: tracyyun | last post by:
Dear forum friends, With the development of smart home technology, a variety of wireless communication protocols have appeared on the market, such as Zigbee, Z-Wave, Wi-Fi, Bluetooth, etc. Each protocol has its own unique characteristics and advantages, but as a user who is planning to build a smart home system, I am a bit confused by the choice of these technologies. I'm particularly interested in Zigbee because I've heard it does some...
0
7309
agi2029
by: agi2029 | last post by:
Let's talk about the concept of autonomous AI software engineers and no-code agents. These AIs are designed to manage the entire lifecycle of a software development project—planning, coding, testing, and deployment—without human intervention. Imagine an AI that can take a project description, break it down, write the code, debug it, and then launch it, all on its own.... Now, this would greatly impact the work of software developers. The idea...
1
6162
isladogs
by: isladogs | last post by:
The next Access Europe User Group meeting will be on Wednesday 1 May 2024 starting at 18:00 UK time (6PM UTC+1) and finishing by 19:30 (7.30PM). In this session, we are pleased to welcome a new presenter, Adolph Dupré who will be discussing some powerful techniques for using class modules. He will explain when you may want to use classes instead of User Defined Types (UDT). For example, to manage the data in unbound forms. Adolph will...
0
5627
by: conductexam | last post by:
I have .net C# application in which I am extracting data from word file and save it in database particularly. To store word all data as it is I am converting the whole word file firstly in HTML and then checking html paragraph one by one. At the time of converting from word file to html my equations which are in the word document file was convert into image. Globals.ThisAddIn.Application.ActiveDocument.Select();...
0
4150
by: TSSRALBI | last post by:
Hello I'm a network technician in training and I need your help. I am currently learning how to create and manage the different types of VPNs and I have a question about LAN-to-LAN VPNs. The last exercise I practiced was to create a LAN-to-LAN VPN between two Pfsense firewalls, by using IPSEC protocols. I succeeded, with both firewalls in the same network. But I'm wondering if it's possible to do the same thing, with 2 Pfsense firewalls...

By using Bytes.com and it's services, you agree to our Privacy Policy and Terms of Use.

To disable or enable advertisements and analytics tracking please visit the manage ads & tracking page.