Re: Query execution plan

[Rick James]

> PRIMARY KEY (`time`)
Risky. You could get two records with the same time. (But I don't know what kind of clock has the resolution you seem to have.)

> reason for not having one big table is speed of recalling data
Have you demonstrated that it makes any difference?

Do this:
EXPLAIN PARTITIONS select * from table_288 where time <= 5706114878826086400 order by time desc limit 1 ;
Do it for both a "range" query and an "index" query.

> This query runs very fast on the majority of tables, but very slow on a few.
Provide more details, so we can discuss what is different.

> 1 CEV0L05RT.MIN 93 range
Are you saying that you have a 93-row table partitioned?

> 4,242 ILM1H03B 111,476,494 index

When it decides to be "range" is the time (5706114878826086400, or whatever) in the table? I am wondering if this leads to the diff.
Is the query just like the SELECT, above?