Mysql – How to make a FULLTEXT search with ORDER BY fast

full-text-searchmariadbMySQLmysql-5.6

I'm trying to get a simple FULLTEXT match to be faster when using order by on another column on a table with over 100 million rows. The basis is one table with a fulltext on two columns and I want to search the database but order it by either the primary (least/most recent) or popularity. Is it possible to make a FULLTEXT with an order by on another indexed column fast? SQL Fiddle below with schema and explains of all queries:

See SQL Fiddle #1

What's very fast so far is denormalization of search columns in a separate table and a join but I would rather not have another table if not necessary. SQL Fiddle below (denormalized query at the end):

See SQL Fiddle #2

Best Answer

The first problem here is something you cannot control. What is it ???

The Query Optimizer's reaction to a FULLTEXT index. Why ?

(Now imagine the start of a STAR WARS movie with scrolling words...)

A very long time ago, I discovered that FULLTEXT indexes will nullify the MySQL Query Optimizer's use of other indexes

Jun 10, 2011 : Mysql fulltext boolean ignore phrase
Oct 25, 2011 : FULLTEXT index ignored in BOOLEAN MODE with 'number of words' conditional
Mar 18, 2012 : Why is LIKE more than 4x faster than MATCH...AGAINST on a FULLTEXT index in MySQL?
AND MORE LIKE IT !!!

SUGGESTION

Refactor the Query so that the MATCH ...AGAINST collects keys only

EXAMPLE #1

SELECT a.id FROM a 
WHERE
MATCH (`search1`,`search2`) AGAINST ('aaaa' IN BOOLEAN MODE)
ORDER BY a.id DESC
LIMIT 5;

should become something like

SELECT N.id FROM
(SELECT id FROM a WHERE
MATCH (`search1`,`search2`) AGAINST ('aaaa' IN BOOLEAN MODE)) M
INNER JOIN a N USING (id)
ORDER BY N.id DESC LIMIT 5;

EXAMPLE #2

SELECT a.id,a.popularity FROM a 
WHERE
MATCH (`search1`,`search2`) AGAINST ('aaaa' IN BOOLEAN MODE)
ORDER BY a.popularity DESC
LIMIT 5;

should become something like

SELECT N.id,N.popularity FROM
(SELECT id FROM a WHERE
MATCH (`search1`,`search2`) AGAINST ('aaaa' IN BOOLEAN MODE)) M
INNER JOIN a N USING (id)
ORDER BY N.popularity DESC LIMIT 5;

CONCLUSION

The main idea: Collect the keys using MATCH ...AGAINST and join it back to the source table

Related Solutions

Sql-server – SQL Server Fulltext search against big amount of search terms over some period of time

Execution plan XML would be useful here but as a shot in the dark:

Daft as it may sound, try pushing either the [id] or date predicate into the CONTAINSTABLE query. See SQL Server 2005 Full-Text Queries on Large Catalogs: Lessons Learned - Consider embedding filter conditions as keywords in the indexed text.
I don't remember where but I recall reading an article or blog post some time ago that flagged big OR full-text queries as problematic. IIRC the suggested hack/workaround was to issue a UNION query instead.

For your example, the UNION for 2) would be along the lines of:

INSERT
    @processed_rules
    (
    rule_id
    , story_id
    )
SELECT
    @rule_id
  , s.[key]
FROM
   CONTAINSTABLE(Stories, (header, body), '"term one"')

UNION

SELECT
    @rule_id
  , s.[key]
FROM
   CONTAINSTABLE(Stories, (header, body), '"term two"')

UNION

  SELECT
    @rule_id
  , s.[key]
FROM
   CONTAINSTABLE(Stories, (header, body), '"term three"')

Mysql – why/how does the number of matched columns influences the way of excecuting a query

I can provide with a general explanation, but it may not apply specifically to your particular case:

The way decision making works is by evaluation cost of execution plan, then picking up what is hopefully the cheapest plan. This you already know.

When it comes to indexing, though, stuff are getting interesting. The way to evaluate the usefulness or viability of an index is to estimate the selectivity given some value.

For the moment, forget about your FULLTEXT index, and let's assume a simple index on some column col1, and another index on some column col2. Given the following two queries:

SELECT * FROM t WHERE col1 < 10 and col2 = 4;
SELECT * FROM t WHERE col1 BETWEEN 100 AND 110 and col2 = 4;

It may happen that the query is evaluated differently in these two cases. Why? Because it may happen that col2 = 4 returns more rows than col1 < 10, in which case we prefer to use index on col1. But then, it may return less rows than col1 BETWEEN 100 AND 110, in which case we prefer the index on col2.

Your case is not very much different. MySQL estimates the number of rows returned by some index query. When you use more columns, MySQL gets the impression your index is likely to result with few rows. So it chooses to start with TableA, then joins what should be very few rows with TableB.

But if MySQL believes the index to return many rows, it may prefer starting with TableB. Why is that? Because you are sorting on indexed columns of TableB. Sorting is a lot of work, too. So MySQL may choose to first sort the rows, then join to TableA and filter by fulltext index. It may not be a bad idea if the fulltext search yields with many rows anyhow.