Shulou(Shulou.com)06/01 Report--

| | 0 | SELECT STATEMENT | | 107 | 2782 | 3 (34) | 00:00:01 |

| | 1 | PX COORDINATOR |

| | 2 | PX SEND QC (RANDOM) |: TQ10001 | 107 | 2782 | 3 (34) | 00:00:01 | Q1Power01 | P-> S | QC (RAND) |

| | 3 | HASH GROUP BY | | 107 | 2782 | 3 (34) | 00:00:01 | Q1Power01 | PCWP |

| | 4 | PX RECEIVE | | 107 | 2782 | 3 (34) | 00:00:01 | Q1Power01 | PCWP |

| | 5 | PX SEND HASH |: TQ10000 | 107 | 2782 | 3 (34) | 00:00:01 | Q1Power00 | P-> P | HASH |

| | 6 | HASH GROUP BY | | 107 | 2782 | 3 (34) | 00:00:01 | Q1Power00 | PCWP |

| | 7 | PX BLOCK ITERATOR | | 107 | 2782 | 2 (0) | 00:00:01 | Q1Power00 | PCWC |

| | 8 | TABLE ACCESS FULL | TEST | 107 | 2782 | 2 (0) | 00:00:01 | Q1Power00 | PCWP |

Perform the steps:

1. The TEST table of the parallel service process is scanned.

2. The parallel service process accesses the data block in ITERATOR (iterative) way, that is, the parallel coordination process distributes a data slice to each parallel service process, on this data chip, the parallel service process accesses each data block (Iterator) sequentially, and all parallel service processes transmit the scanned data blocks to another group of parallel service processes (parent processes) for Hash Group operations.

3. The parallel service parent process does Hash Group operations on the data passed by the child process.

4. The parallel service process (child process) sends out the processed data.

5. The parallel service process (parent process) receives the processed data.

6. Merge the processed data and send them to the parallel coordination process (QC:Query Conordinator) in a random order.

7. The parallel coordination process sends the processing results to the user.

Parallel mechanism: | @ | first, Oracle creates a process to coordinate the transmission of information between parallel service processes. The coordination process divides the dataset that needs to be operated (such as the data blocks of the table) into many parts, called parallel processing units, and then the parallel coordination process assigns a data unit to each parallel process. | @ | | @ | for example, if there are four parallel service processes, they will process their assigned units at the same time. When a parallel service process is finished, the coordination process will assign them another unit, and so on. Until all the data on the table is processed, the coordination process is responsible for merging each small collection into a large set as the final execution result, which is returned to the user. | @ | | @ | parallel processing mechanism actually divides a data set to be scanned into many small data sets. Oracle starts several parallel service processes to process these small data sets at the same time, and finally summarizes these results and returns them to the user as the final processing result. | @ | | @ | this method of parallel data processing is very useful in OLAP systems. Tables in OLAP systems are usually very large. If the system has a large number of CPU, let all CPU process the data together, the effect will be much higher than serial execution. | @ | | @ | however, generally speaking, parallelism is not suitable for OLTP systems, because in almost all SQL operations on OLTP systems, the data access path is basically index access, and the returned result set is very small. The processing speed of such SQL operations is generally very fast, so there is no need to enable parallelism.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.