Using pipelines to combine operations

In a DSS workflow, SQL pipeline is the process of mixing multiple subsequent recipes, each using the same SQL engine. You can then execute a single job activity containing these integrated recipes (possibly visual recipes and SQL Query recipes).

Typically, SQL queries are converted into expressions in relational algebra, that is, a series of relational operations. If we only performed one operation at a time, it would incur too much cost because we would need to create temporary files on the disc to store the results of these temporary operations. Large temporary files must be created and stored on the disc, but this takes time and is usually not required since the following process will use these files immediately as input. It is common practice to generate query execution code that corresponds to algorithms for specific combinations of operations in the query to reduce the number of temporary files.

Example

For example, a method with two input files and one output file could perform a JOIN and two SELECT operations on the input files, and a final PROJECT operation on the result file, rather than implementing each of these operations individually . Instead of creating four temporary files, we apply the method and get only one resulting file. Pipelining or stream-based processing are the terms used for this purpose.

in conclusion

In order to perform a large number of operations, it is common practice to dynamically generate query execution code. Queries are generated by creating code that contains numerous algorithms corresponding to different processes. After an operation is performed, the resulting tuple is generated and used as input for other operations. For example, if two SELECT operations on a base relationship are followed by a JOIN operation, the tuples created by each SELECT operation will be fed into a stream or pipeline and used as input to the JOIN process.

The above is the detailed content of Using pipelines to combine operations. For more information, please follow other related articles on the PHP Chinese website!