I have searched for a long time (over 35 days) before write here.

I optimised my mysql database as much as I can but unfortunately in configuration part I could not go further. Website is a classified ads service. It runs on php7.4 (laravel framework), I am using memcached and mysql.

Mysql Server version: 5.7.26-0ubuntu0.16.04.1 (Ubuntu) RAM 16GB Cores: 6 SSD 400GB Server type: VPS Table type is InnoDB

When I run like query with FORCE INDEX it takes between 40ms-78ms (some times 100ms - rare), without FORE INDEX it takes 95ms.

select * 
from `table_name` 
where `access`=1 
     and ((`title` LIKE '%alabama%' or `about` LIKE '%alabama%') 
       or (`title` LIKE '%rental%' or `about` LIKE '%rental%') 
       or (`title` LIKE '%house%' or `about` LIKE '%house%') 
       or (`title` LIKE '%cheap%' or `about` LIKE '%cheap%'))
# result 95.3ms / 2063 rows - total rows ~194.849

select `table_name`.`category_id`, COUNT(table_name.id) AS total_count 
from `table_name` 
where (`table_name`.`access` = 1) 
      and ((`title` LIKE '%alabama%' or `about` LIKE '%alabama%') 
        or (`title` LIKE '%rental%' or `about` LIKE '%rental%') 
        or (`title` LIKE '%house%' or `about` LIKE '%house%') 
        or (`title` LIKE '%cheap%' or `about` LIKE '%cheap%')) 
group by `table_name`.`category_id`;
# result 3.51s / 159 rows

> show variables;
auto_increment_increment    1
auto_increment_offset   1
autocommit  ON
automatic_sp_privileges ON
avoid_temporal_upgrade  OFF
back_log    80
basedir /usr/
big_tables  OFF
binlog_cache_size   32768
binlog_checksum CRC32
binlog_direct_non_transactional_updates OFF
binlog_error_action ABORT_SERVER
binlog_format   ROW
binlog_group_commit_sync_delay  0
binlog_group_commit_sync_no_delay_count 0
binlog_gtid_simple_recovery ON
binlog_max_flush_queue_time 0
binlog_order_commits    ON
binlog_row_image    FULL
binlog_rows_query_log_events    OFF
binlog_stmt_cache_size  32768
binlog_transaction_dependency_history_size  25000
binlog_transaction_dependency_tracking  COMMIT_ORDER
block_encryption_mode   aes-128-ecb
bulk_insert_buffer_size 8388608
character_set_client    utf8
character_set_connection    utf8
character_set_database  utf8
character_set_filesystem    binary
character_set_results   utf8
character_set_server    latin1
character_set_system    utf8
character_sets_dir  /usr/share/mysql/charsets/
check_proxy_users   OFF
collation_connection    utf8_general_ci
collation_database  utf8_general_ci
collation_server    latin1_swedish_ci
completion_type NO_CHAIN
concurrent_insert   AUTO
connect_timeout 10
core_file   OFF
datadir /var/lib/mysql/
date_format %Y-%m-%d
datetime_format %Y-%m-%d %H:%i:%s
default_authentication_plugin   mysql_native_password
default_password_lifetime   0
default_storage_engine  InnoDB
default_tmp_storage_engine  InnoDB
default_week_format 0
delay_key_write ON
delayed_insert_limit    100
delayed_insert_timeout  300
delayed_queue_size  1000
disconnect_on_expired_password  ON
div_precision_increment 4
end_markers_in_json OFF
enforce_gtid_consistency    OFF
eq_range_index_dive_limit   200
error_count 0
event_scheduler OFF
expire_logs_days    10
explicit_defaults_for_timestamp OFF
flush   OFF
flush_time  0
foreign_key_checks  ON
ft_boolean_syntax   + -><()~*:""&|
ft_max_word_len 84
ft_min_word_len 4
ft_query_expansion_limit    20
ft_stopword_file    (built-in)
general_log OFF
group_concat_max_len    1024
gtid_executed_compression_period    1000
gtid_mode   OFF
gtid_next   AUTOMATIC
have_compress   YES
have_crypt  YES
have_dynamic_loading    YES
have_geometry   YES
have_openssl    DISABLED
have_profiling  YES
have_query_cache    YES
have_rtree_keys YES
have_ssl    DISABLED
have_statement_timeout  YES
have_symlink    YES
host_cache_size 279
identity    0
ignore_builtin_innodb   OFF
innodb_adaptive_flushing    ON
innodb_adaptive_flushing_lwm    10
innodb_adaptive_hash_index  ON
innodb_adaptive_hash_index_parts    8
innodb_adaptive_max_sleep_delay 150000
innodb_api_bk_commit_interval   5
innodb_api_disable_rowlock  OFF
innodb_api_enable_binlog    OFF
innodb_api_enable_mdl   OFF
innodb_api_trx_level    0
innodb_autoextend_increment 64
innodb_autoinc_lock_mode    1
innodb_buffer_pool_chunk_size   134217728
innodb_buffer_pool_dump_at_shutdown ON
innodb_buffer_pool_dump_now OFF
innodb_buffer_pool_dump_pct 25
innodb_buffer_pool_filename ib_buffer_pool
innodb_buffer_pool_instances    1
innodb_buffer_pool_load_abort   OFF
innodb_buffer_pool_load_at_startup  ON
innodb_buffer_pool_load_now OFF
innodb_buffer_pool_size 134217728
innodb_change_buffer_max_size   25
innodb_change_buffering all
innodb_checksum_algorithm   crc32
innodb_checksums    ON
innodb_cmp_per_index_enabled    OFF
innodb_commit_concurrency   0
innodb_compression_failure_threshold_pct    5
innodb_compression_level    6
innodb_compression_pad_pct_max  50
innodb_concurrency_tickets  5000
innodb_data_file_path   ibdata1:12M:autoextend
innodb_deadlock_detect  ON
innodb_default_row_format   dynamic
innodb_disable_sort_file_cache  OFF
innodb_doublewrite  ON
innodb_fast_shutdown    1
innodb_file_format  Barracuda
innodb_file_format_check    ON
innodb_file_format_max  Barracuda
innodb_file_per_table   ON
innodb_fill_factor  100
innodb_flush_log_at_timeout 1
innodb_flush_log_at_trx_commit  1
innodb_flush_neighbors  1
innodb_flush_sync   ON
innodb_flushing_avg_loops   30
innodb_force_load_corrupted OFF
innodb_force_recovery   0
innodb_ft_cache_size    8000000
innodb_ft_enable_diag_print OFF
innodb_ft_enable_stopword   ON
innodb_ft_max_token_size    84
innodb_ft_min_token_size    3
innodb_ft_num_word_optimize 2000
innodb_ft_result_cache_limit    2000000000
innodb_ft_sort_pll_degree   2
innodb_ft_total_cache_size  640000000
innodb_io_capacity  200
innodb_io_capacity_max  2000
innodb_large_prefix ON
innodb_lock_wait_timeout    50
innodb_locks_unsafe_for_binlog  OFF
innodb_log_buffer_size  16777216
innodb_log_checksums    ON
innodb_log_compressed_pages ON
innodb_log_file_size    50331648
innodb_log_files_in_group   2
innodb_log_group_home_dir   ./
innodb_log_write_ahead_size 8192
innodb_lru_scan_depth   1024
innodb_max_dirty_pages_pct  75.000000
innodb_max_dirty_pages_pct_lwm  0.000000
innodb_max_purge_lag    0
innodb_max_purge_lag_delay  0
innodb_max_undo_log_size    1073741824
innodb_numa_interleave  OFF
innodb_old_blocks_pct   37
innodb_old_blocks_time  1000
innodb_online_alter_log_max_size    134217728
innodb_open_files   431
innodb_optimize_fulltext_only   OFF
innodb_page_cleaners    1
innodb_page_size    16384
innodb_print_all_deadlocks  OFF
innodb_purge_batch_size 300
innodb_purge_rseg_truncate_frequency    128
innodb_purge_threads    4
innodb_random_read_ahead    OFF
innodb_read_ahead_threshold 56
innodb_read_io_threads  4
innodb_read_only    OFF
innodb_replication_delay    0
innodb_rollback_on_timeout  OFF
innodb_rollback_segments    128
innodb_sort_buffer_size 1048576
innodb_spin_wait_delay  6
innodb_stats_auto_recalc    ON
innodb_stats_include_delete_marked  OFF
innodb_stats_method nulls_equal
innodb_stats_on_metadata    OFF
innodb_stats_persistent ON
innodb_stats_persistent_sample_pages    20
innodb_stats_sample_pages   8
innodb_stats_transient_sample_pages 8
innodb_status_output    OFF
innodb_status_output_locks  OFF
innodb_strict_mode  ON
innodb_support_xa   ON
innodb_sync_array_size  1
innodb_sync_spin_loops  30
innodb_table_locks  ON
innodb_temp_data_file_path  ibtmp1:12M:autoextend
innodb_thread_concurrency   0
innodb_thread_sleep_delay   10000
innodb_undo_directory   ./
innodb_undo_log_truncate    OFF
innodb_undo_logs    128
innodb_undo_tablespaces 0
innodb_use_native_aio   ON
innodb_version  5.7.26
innodb_write_io_threads 4
insert_id   0
interactive_timeout 28800
internal_tmp_disk_storage_engine    InnoDB
join_buffer_size    262144
keep_files_on_create    OFF
key_buffer_size 16777216
key_cache_age_threshold 300
key_cache_block_size    1024
key_cache_division_limit    100
keyring_operations  ON
large_files_support ON
large_page_size 0
large_pages OFF
last_insert_id  0
lc_messages en_US
lc_time_names   en_US
license GPL
local_infile    ON
lock_wait_timeout   31536000
locked_in_memory    OFF
log_bin OFF
log_bin_trust_function_creators OFF
log_bin_use_v1_row_events   OFF
log_builtin_as_identified_by_password   OFF
log_error_verbosity 3
log_output  FILE
log_queries_not_using_indexes   OFF
log_slave_updates   OFF
log_slow_admin_statements   OFF
log_slow_slave_statements   OFF
log_statements_unsafe_for_binlog    ON
log_syslog  OFF
log_syslog_facility daemon
log_syslog_include_pid  ON
log_throttle_queries_not_using_indexes  0
log_timestamps  UTC
log_warnings    2
long_query_time 10.000000
low_priority_updates    OFF
lower_case_file_system  OFF
lower_case_table_names  0
master_info_repository  FILE
master_verify_checksum  OFF
max_allowed_packet  16777216
max_binlog_cache_size   18446744073709547520
max_binlog_size 104857600
max_binlog_stmt_cache_size  18446744073709547520
max_connect_errors  100
max_connections 151
max_delayed_threads 20
max_digest_length   1024
max_error_count 64
max_execution_time  0
max_heap_table_size 16777216
max_insert_delayed_threads  20
max_join_size   18446744073709551615
max_length_for_sort_data    1024
max_points_in_geometry  65536
max_prepared_stmt_count 16382
max_relay_log_size  0
max_seeks_for_key   18446744073709551615
max_sort_length 1024
max_sp_recursion_depth  0
max_tmp_tables  32
max_user_connections    0
max_write_lock_count    18446744073709551615
metadata_locks_cache_size   1024
metadata_locks_hash_instances   8
min_examined_row_limit  0
multi_range_count   256
myisam_data_pointer_size    6
myisam_max_sort_file_size   9223372036853727232
myisam_mmap_size    18446744073709551615
myisam_recover_options  BACKUP
myisam_repair_threads   1
myisam_sort_buffer_size 8388608
myisam_stats_method nulls_unequal
myisam_use_mmap OFF
mysql_native_password_proxy_users   OFF
net_buffer_length   16384
net_read_timeout    30
net_retry_count 10
net_write_timeout   60
new OFF
ngram_token_size    2
offline_mode    OFF
old OFF
old_alter_table OFF
old_passwords   0
open_files_limit    1024
optimizer_prune_level   1
optimizer_search_depth  62
optimizer_switch    index_merge=on,index_merge_union=on,index_merge_sort_union=on,index_merge_intersection=on,engine_condition_pushdown=on,index_condition_pushdown=on,mrr=on,mrr_cost_based=on,block_nested_loop=on,batched_key_access=off,materialization=on,semijoin=on,loosescan=on,firstmatch=on,duplicateweedout=on,subquery_materialization_cost_based=on,use_index_extensions=on,condition_fanout_filter=on,derived_merge=on
optimizer_trace enabled=off,one_line=off
optimizer_trace_features    greedy_search=on,range_optimizer=on,dynamic_range=on,repeated_subselect=on
optimizer_trace_limit   1
optimizer_trace_max_mem_size    16384
optimizer_trace_offset  -1
parser_max_mem_size 18446744073709551615
performance_schema  ON
performance_schema_accounts_size    -1
performance_schema_digests_size 5000
performance_schema_events_stages_history_long_size  1000
performance_schema_events_stages_history_size   10
performance_schema_events_statements_history_long_size  1000
performance_schema_events_statements_history_size   10
performance_schema_events_transactions_history_long_size    1000
performance_schema_events_transactions_history_size 10
performance_schema_events_waits_history_long_size   1000
performance_schema_events_waits_history_size    10
performance_schema_hosts_size   -1
performance_schema_max_cond_classes 80
performance_schema_max_cond_instances   -1
performance_schema_max_digest_length    1024
performance_schema_max_file_classes 80
performance_schema_max_file_handles 32768
performance_schema_max_file_instances   -1
performance_schema_max_index_stat   -1
performance_schema_max_memory_classes   320
performance_schema_max_metadata_locks   -1
performance_schema_max_mutex_classes    210
performance_schema_max_mutex_instances  -1
performance_schema_max_prepared_statements_instances    -1
performance_schema_max_program_instances    -1
performance_schema_max_rwlock_classes   50
performance_schema_max_rwlock_instances -1
performance_schema_max_socket_classes   10
performance_schema_max_socket_instances -1
performance_schema_max_sql_text_length  1024
performance_schema_max_stage_classes    150
performance_schema_max_statement_classes    193
performance_schema_max_statement_stack  10
performance_schema_max_table_handles    -1
performance_schema_max_table_instances  -1
performance_schema_max_table_lock_stat  -1
performance_schema_max_thread_classes   50
performance_schema_max_thread_instances -1
performance_schema_session_connect_attrs_size   512
performance_schema_setup_actors_size    -1
performance_schema_setup_objects_size   -1
performance_schema_users_size   -1
port    3306
preload_buffer_size 32768
profiling   OFF
profiling_history_size  15
protocol_version    10
pseudo_slave_mode   OFF
pseudo_thread_id    591161
query_alloc_block_size  8192
query_cache_limit   1048576
query_cache_min_res_unit    4096
query_cache_size    16777216
query_cache_type    OFF
query_cache_wlock_invalidate    OFF
query_prealloc_size 8192
rand_seed1  0
rand_seed2  0
range_alloc_block_size  4096
range_optimizer_max_mem_size    8388608
rbr_exec_mode   STRICT
read_buffer_size    131072
read_only   OFF
read_rnd_buffer_size    262144
relay_log_info_file relay-log.info
relay_log_info_repository   FILE
relay_log_purge ON
relay_log_recovery  OFF
relay_log_space_limit   0
report_port 3306
require_secure_transport    OFF
rpl_stop_slave_timeout  31536000
secure_auth ON
server_id   0
server_id_bits  32
server_uuid dd1352c2-4962-11e7-982a-0050563d1bfa
session_track_gtids OFF
session_track_schema    ON
session_track_state_change  OFF
session_track_system_variables  time_zone,autocommit,character_set_client,character_set_results,character_set_connection
session_track_transaction_info  OFF
sha256_password_proxy_users OFF
show_compatibility_56   OFF
show_create_table_verbosity OFF
show_old_temporals  OFF
skip_external_locking   ON
skip_name_resolve   OFF
skip_networking OFF
skip_show_database  OFF
slave_allow_batching    OFF
slave_checkpoint_group  512
slave_checkpoint_period 300
slave_compressed_protocol   OFF
slave_exec_mode STRICT
slave_load_tmpdir   /tmp
slave_max_allowed_packet    1073741824
slave_net_timeout   60
slave_parallel_type DATABASE
slave_parallel_workers  0
slave_pending_jobs_size_max 16777216
slave_preserve_commit_order OFF
slave_rows_search_algorithms    TABLE_SCAN,INDEX_SCAN
slave_skip_errors   OFF
slave_sql_verify_checksum   ON
slave_transaction_retries   10
slow_launch_time    2
slow_query_log  OFF
sort_buffer_size    262144
sql_auto_is_null    OFF
sql_big_selects ON
sql_buffer_result   OFF
sql_log_bin ON
sql_log_off OFF
sql_notes   ON
sql_quote_show_create   ON
sql_safe_updates    OFF
sql_select_limit    18446744073709551615
sql_slave_skip_counter  0
sql_warnings    OFF
stored_program_cache    256
super_read_only OFF
sync_binlog 1
sync_frm    ON
sync_master_info    10000
sync_relay_log  10000
sync_relay_log_info 10000
system_time_zone    CET
table_definition_cache  615
table_open_cache    431
table_open_cache_instances  16
thread_cache_size   8
thread_handling one-thread-per-connection
thread_stack    196608
time_format %H:%i:%s
time_zone   SYSTEM
timestamp   1612972467.547293
tls_version TLSv1,TLSv1.1
tmp_table_size  16777216
tmpdir  /tmp
transaction_alloc_block_size    8192
transaction_allow_batching  OFF
transaction_isolation   REPEATABLE-READ
transaction_prealloc_size   4096
transaction_read_only   OFF
transaction_write_set_extraction    OFF
tx_isolation    REPEATABLE-READ
tx_read_only    OFF
unique_checks   ON
updatable_views_with_limit  YES
version 5.7.26-0ubuntu0.16.04.1
version_comment (Ubuntu)
version_compile_machine x86_64
version_compile_os  Linux
wait_timeout    28800
warning_count   0


EXPLAIN select `table_name`.`category_id`, COUNT(table_name.id) AS total_count 
from `table_name` 
where (`table_name`.`access` = 1) 
      and ((`title` LIKE '%alabama%' or `about` LIKE '%alabama%') 
        or (`title` LIKE '%rental%' or `about` LIKE '%rental%') 
        or (`title` LIKE '%house%' or `about` LIKE '%house%') 
        or (`title` LIKE '%cheap%' or `about` LIKE '%cheap%')) 
group by `table_name`.`category_id`;

| id | select_type | table   | partitions | type | possible_keys      | key    | key_len | ref   | rows  | filtered | Extra                                                               |
|  1 | SIMPLE      | table_name | NULL       | ref  | category_id,access | access | 2       | const | 23916 |    61.02 | Using index condition; Using where; Using temporary; Using filesort |
1 row in set, 1 warning (0.00 sec)

EXPLAIN select * 
from `table_name` 
where `access`=1 
     and ((`title` LIKE '%alabama%' or `about` LIKE '%alabama%') 
       or (`title` LIKE '%rental%' or `about` LIKE '%rental%') 
       or (`title` LIKE '%house%' or `about` LIKE '%house%') 
       or (`title` LIKE '%cheap%' or `about` LIKE '%cheap%'))
| id | select_type | table   | partitions | type | possible_keys | key    | key_len | ref   | rows  | filtered | Extra       |
|  1 | SIMPLE      | table_name | NULL       | ref  | access        | access | 2       | const | 23916 |    61.02 | Using where |
1 row in set, 1 warning (0.02 sec)


| table_name | CREATE TABLE `table_name` (
  `category_id` int(11) NOT NULL,
  `url` text NOT NULL,
  `keywords` varchar(300) DEFAULT NULL,
  `country` int(11) NOT NULL,
  `type` int(11) NOT NULL,
  `price` varchar(250) NOT NULL,
  `currency` int(11) NOT NULL,
  `title` varchar(500) CHARACTER SET utf8mb4 NOT NULL DEFAULT '',
  `about` text CHARACTER SET utf8mb4 NOT NULL,
  `text_combained` text CHARACTER SET utf8mb4,
  `has_images` tinyint(1) DEFAULT '0',
  `images_count` smallint(2) DEFAULT '0',
  `new` tinyint(1) DEFAULT '1',
  `person` int(11) DEFAULT NULL,
  `fullname` varchar(255) NOT NULL,
  `email` varchar(255) NOT NULL,
  `email_approval_token` varchar(255) DEFAULT NULL,
  `email_approved_at` timestamp NULL DEFAULT NULL,
  `contact` varchar(255) NOT NULL,
  `whatsapp` varchar(255) DEFAULT NULL,
  `period` int(11) NOT NULL,
  `period_tmps` int(11) NOT NULL,
  `access` tinyint(1) DEFAULT '0',
  `status` smallint(2) DEFAULT NULL,
  `address` varchar(255) DEFAULT NULL,
  `user_id` int(11) NOT NULL,
  `timestamp` int(11) NOT NULL,
  `views` int(11) NOT NULL DEFAULT '0',
  `ip` varchar(64) DEFAULT NULL,
  `xml` tinyint(1) DEFAULT NULL,
  `premium` int(11) DEFAULT NULL,
  `date` varchar(20) NOT NULL,
  `time` varchar(20) NOT NULL,
  `datetime` datetime NOT NULL,
  `is_published` tinyint(1) DEFAULT NULL,
  `is_deleted` tinyint(1) DEFAULT NULL,
  `created_at` datetime DEFAULT NULL,
  `published_at` datetime DEFAULT NULL,
  `updated_at` datetime DEFAULT NULL,
  `deleted_at` datetime DEFAULT NULL,
  PRIMARY KEY (`id`),
  KEY `category_id` (`category_id`),
  KEY `country` (`country`),
  KEY `access` (`access`),
  KEY `user_id` (`user_id`)


| Table   | Non_unique | Key_name    | Seq_in_index | Column_name | Collation | Cardinality | Sub_part | Packed | Null | Index_type | Comment | Index_comment |
| table_name |          0 | PRIMARY     |            1 | id          | A         |      194835 |     NULL | NULL   |      | BTREE      |         |               |
| table_name |          1 | category_id |            1 | category_id | A         |         414 |     NULL | NULL   |      | BTREE      |         |               |
| table_name |          1 | country     |            1 | country     | A         |          71 |     NULL | NULL   |      | BTREE      |         |               |
| table_name |          1 | access      |            1 | access      | A         |          10 |     NULL | NULL   | YES  | BTREE      |         |               |
| table_name |          1 | user_id     |            1 | user_id     | A         |        5168 |     NULL | NULL   |      | BTREE      |         |               |

enter image description here

  • 5
    Provide the table definition and the query/execution plan as well as the exact queries you're running
    – bbaird
    Feb 8 at 21:01
  • Assuming table.access =1 one accounts for a large portion of the table its unlikely to help indexing that. The rest of your query cannot use an index (without a full scan). Look at EXPLAIN {query} on your two options. If your set of search critiera is fixed consider mapping those to a primary key in a second table.
    – danblack
    Feb 9 at 3:19
  • @bbaird Thank you for your comment. I have updated question add more details and table information, main query what takes long time. Feb 9 at 20:26
  • @danblack I have shared EXPLAIN queries Feb 9 at 20:26
  • avoid using OR in queries.. They are pretty slow and consumes alot of resources
    – clarkk
    Feb 10 at 18:56

The usual index type (btree) on text columns orders values alphabetically, so it is fast to find a specific value, or a range between two bounds (say, A to B), or values that start with a specific prefix. This means it will optimize LIKE 'foo%' quite well.

It's a bit like a dictionary. You can quickly find a word if you know the first letters, but the btree index provides no help at all for queries like '%foo%'. It's like looking for all words that contain 'foo' in the dictionary. You'll have to read the whole book.

So... if you want to do fulltext search, use a fulltext index. It uses a different indexing algorithm, which takes into account each word in the text. So it will optimize "does the text contain the word 'house'" which in your case is much faster than the btree indew, which will optimize "does the text start with 'house'".

  • Thank you for your replay. I had queries as you explained, but results were not much as I expected. FULLTEXT search worked quite fast but unfortunately with website language it is not successful especially I had troubles with numbers (exp. 1 room rental house). Feb 9 at 20:42
  • Yes it will not index "words" of less than 3 charachers, also numbers I think. For this you would need smarter fulltext engine, with natural language processing, so it knows that in "1 room" the "1" has a special meaning when it is next to "room"...
    – bobflux
    Feb 9 at 22:59
  • Usually classified sites solve this issue by having a few mandatory fields like surface, rooms, zipcode, type of building, floor, etc. These can be indexed independently and then you no longer need fulltext search. In fact I've never seen such a site using fulltext search. Of course the user who submits the ad must enter correct values, but it's easy to motivate them, as entering correct values means more people will find their ad and buy what they are selling.
    – bobflux
    Feb 10 at 14:51
  • Thank you for replay. Honestly, website not only for real estate it is mixed, cars, animals, housing, fishing, hobby etc. people prefer to search and find what they desire. Another big issue is language, if it was in english it would be easy to use FT. Feb 10 at 15:42
  • Have you tried a dedicated fulltext search that supports your language?
    – bobflux
    Feb 10 at 16:41

The villain is not GROUP BY and COUNT, but LIKE "%..." and OR. But, I will launch into a major rewrite of your application...

In order to make an efficient real estate search tool, you must

  • Divide up the data into "text search" and "numeric search"
  • Also split into "commonly searched on" versus "rarely searched on".

Common numeric things are price range and number of bedrooms. Things like "2 dishwashers" should not be included, because of the rarety.

Things like "gazebo" are very rare, hence very selective. They are best handled in a FULLTEXT search.

Further, the app should gather the information presented by the user and carefully direct it to where it can be searched. In particular, all things that are best done by FT should be gathered into one MATCH. And there may as well be a single TEXT column where "gazebo", etc might occur.

As you are building the SELECT, sometimes there will be only a FULLTEXT search, sometimes tests against numeric columns, sometimes both. Note: The FT part of the WHERE clause will be performed first if present, then further filtering would happen with the other tests.

Hence, ... A search only for "gazebo" will be quite efficient. Meanwhile, things like "3 bedrooms AND price between ..." (but not FT test) will need some other techniques.

So, let's discuss searching on multiple numeric columns...

A "composite" INDEX is one that has multiple columns. It will work well if the first column(s) mentioned are tested with = and then one "range" column can be used. So, for the example above, INDEX(bedrooms, price) is optimal and, depending on the values, may narrow down the search nicely.

Unfortunately, multiple range tests is very hard to optimize. That is "2-3 bedrooms and at least 1.5 bathc and price < 1234" cannot be performed efficiently.

"2-3 bedrooms" can be phrased as "2 bedrooms OR 3 bedrooms", but that only pushes the problem into the inefficiency of "OR". However, that can be optimized by a UNION:

( SELECT ... WHERE bedrooms = 2 AND price ... )
( SELECT ... WHERE bedrooms = 3 AND price ... )

In this example, each SELECT can efficiently use INDEX(bedrooms, price), leading to a "small" number of rows, then the UNION combines the results. (Note ALL is faster than DISTINCT, but is relevant only because the two sets of rows do not overlay because of a different number of bedrooms.)

Notice how I am pushing more and more of the programming effort away from SQL and into the application. Sorry; that's a fact of life.

Back to your example -- A fulltext search AGAINST("Alabama rental house cheap") would be very fast and efficient. But you might want to pull "cheap" out of the set of words and turn it into "ORDER BY price ASC", since that is what is needed.

Your dataset might have "... Birmingham Jefferson county Alabama ..." to help cover various bases such as "near Birmingham" or "in Jefferson county".

More FT tips...

"Stopwords" and "short" word will be ignored.

Putting + in front of a word can help or hurt. It can help by saying that this is an important keyword. It can hurt if the user says "cheap" but no title uses that word.

Don't be tempted by EAV, as I discuss here: http://mysql.rjweb.org/doc.php/eav

Sorry, but generic 3rd party software (Laravel, etc) tends to generate sub-optimal SQL; you need hands-on SQL to get the performance you are asking for.

Your 24K rows is relatively small; but LIKE with a leading wildcard will get slower and slower, growing linearly with the size of the dataset.

DESCRIBE is less descriptive than SHOW CREATE TABLE. It fails to show the indexes and the engine, among other things.

How much RAM? What is the setting of innodb_buffer_pool_size?

So, how to do "2 dishwashers"? One approach is to make up a word that represents that (perhaps "more_than_one_dw") and have code that transforms the incoming data into that word and the user queries into that word. Then the FT request will be very fast and efficient.

  • Thanks for valuable knowledge sharing. I already tried FT for some time it worked fine then some troubles start to compe up. For example: FT does not support webiste language (Azerbaijani), people search with different short words, unicode and special characters also was breaking search constantly. Then I returned to LIKE. I update question with more mysql configuration details and SHOW CREATE TABLE. RAM is 16 GB Feb 10 at 16:04
  • @tima.lennon - Add the UTF8 to the question; it is significant to the Question and any Solutions. Possibly even add some Azerbaijani words that are problematic.
    – Rick James
    Feb 10 at 17:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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