Optimization of a slow query with a function in front of a date-type field in a Where condition _ function

When an index is created on a field, the index is invalidated when the column is preceded by a function in the Where condition. Sometimes it can be optimized by overwriting SQL.

Such as:

Select COUNT (*) from T_synclog where To_days (now ())-to_days (Create_time) = 0 and status=0 and entity_type= ' org ';

Infer that the SQL implication is to view the number of record bars that were created today.

Table records have 40多万条. The table does not create an index on the create_time. Test on your own virtual machine, it takes about 30 seconds to execute the query.

Workaround: Create an index and then overwrite SQL (avoid using functions before index fields)

Create INDEX Ind_cre on T_synclog (create_time);
Only check the log records created that day: use
EIP;
SELECT COUNT (*) from T_synclog WHERE create_time BETWEEN curdate () and now () and status=0 ' org ';
Check the log records that were created the day before: use
EIP;
SELECT COUNT (*) from T_synclog WHERE create_time BETWEEN date_sub (Curdate (), INTERVAL N Day) and
Date_sub (Date_sub ( Curdate (), INTERVAL n-1 Day), INTERVAL 1 SECOND) and
status=0 and entity_type= ' org ';

The effect is as follows: