# Reads Data - Cont'd 

![`smart`](ExecuteandRead03_aa_005_20220512.png)

-----------------------------------------------------------
Finally, the query results will be returned to the TiDB server.

The requests that can be pushed to the TiKV server currently include TableScan, IndexScan, Selection, TopN, Limit and PartialAggregation, etc. Other more complex operators still need to be processed on a single TiDB server. Therefore, it will be summarized and processed on the TiDB server. After the summary is completed, the result set is formed. If it is a query operation, the result set (including field list and row data) will be returned to the client.

In the execution of reading, part of the operators are pushed to the TiKV server for execution, and the other part is finally aggregated in the TiDB server. Thus, there are two types of computing tasks, they are Cop Task and Root Task, the specific definitions are as follows:

- Cop Task: is a physical plan that needs to be pushed down to the storage engine (TiKV) for computation, and each TiKV server node that receives the request will do the same operation.

- Root Task: is the physical plan that is computed in the TiDB server.