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.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have a table of IP address ranges and their country, here is the structure:

  `startBlock` int(11) unsigned NOT NULL DEFAULT '0',
  `endBlock` int(11) unsigned NOT NULL DEFAULT '0',
  `code` char(2) NOT NULL DEFAULT '',
  `country` varchar(60) NOT NULL DEFAULT '',
  KEY `startEnd` (`startBlock`,`endBlock`),
  KEY `country` (`country`)

The table is typically used to lookup the country of an IP address. I'm trying to understand why these queries do not always use the 'startEnd' index, as my slow query log is filled with entries like this:

# Query_time: 0  Lock_time: 0  Rows_sent: 1  Rows_examined: 120967
SELECT code FROM geoIP WHERE INET_ATON('') > startBlock AND INET_ATON('') < endBlock;

120,967 is the number of rows in the table, so it's doing a full table scan. Here's the explain for that query:

id  select_type  table  type  possible_keys  key    key_len  ref   rows    Extra
1   SIMPLE       geoIP  ALL   startEnd       NULL   NULL     NULL  120967  Using where

Sometimes the index is used though, here's the explain for another - same query, the only difference is the IP address:

id  select_type  table  type   possible_keys  key       key_len  ref    rows    Extra
1   SIMPLE       geoIP  range  startEnd       startEnd  4        NULL   9579    Using where

both queries return 1 row.

Why does MySQL only sometimes use the index? Is there anything I can do to optimise further?

share|improve this question
See this related question regarding further optimisation. The Q there is for SQL Server but I presume that you could use the same in MySQL. – Martin Smith Jun 17 '12 at 21:27
up vote 1 down vote accepted

The general behavior of the MySQL Query Optimizer is as follows: If less than 5% of the rows of a table has to be traversed in an index, the index is used. Otherwise, the index is dismissed and a full table scan is done.

Now look at the index: (startBlock,endBlock).

Here are two questions

  • How many netblock up to the third octet do you have for startBlock ?
  • How many netblock up to the third octet do you have for endBlock ?

To get a hold of the distribution of netblocks, run this query

    IF(ABS(endNetBlock - startNetBlock) >= 256,1,0) NotWithinSameRange,
    IF(startNetBlock - endNetBlock > 0,1,0) StartRangeAheadOfEndRange
    SELECT COUNT(1) rangeCount,startNetBlock,endNetBlock
            (startBlock - MOD(startBlock,256)) startNetBlock,
            (endBlock - MOD(endtBlock,256)) endNetBlock
        FROM geoIP
    ) AA GROUP BY startNetBlock,endNetBlock
) A

This will give you a general idea of

  • how many netblock combinations you have
  • presence of any invalid ranges

If NotWithinSameRange or StartRangeAheadOfEndRange are 1 in any instance, that range has to be clenaed up. You have to be sure that all netBlocks are properly represented.

If you have StartRangeAheadOfEndRange=1, that is a definite problem.

I hope these suggestions provide food for thought.

share|improve this answer

Your Answer


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.