Problem: A/B table left connection, the count one day qualified data, 7 minutes still not out of the results;
Similar to SQL select count (*) from tabble_a A, table_b B where a. usi_day=B.b usi_day (+) and A.r egin_part=B.r egin_part (+) and a. usi_day="20200321" and A.r egin_part="A"
The screening process:
1. The execution plan saw, partitioning index using the use, the count (*), count (1) what tried useless;
2. Look at the data found on the day of A, B table accords with A condition the amount of data that probably around 11 w, count other date data can be found out, but the day basic around 4 w AB table data;
Current location to the question is:
Once the AB table accords with a condition one day the amount of data of more than 10 w, can't count it out;
But count on March 1 to March 20, the data can count out of a total of 50 w +;
Bosses can optimise the opinion of everyone?
CodePudding user response:
To see a big help