Source: "Bi thing" data stream conversion--percent sampling, row sampling
Percent sampling and row sampling can randomly select a set of data from the data source. Both of these tasks can produce two sets of outputs, one randomly selected and the other not selected. These selected data can be sent to the development or test server. The most appropriate application for this task is to build a data mining model and then use these sampled data to validate the model.
To edit this task, select the number of rows or percentages to extract. Percent sampling randomly selects data by percentage from the data source, and row sampling randomly selects the exact number of rows from the data source. You can name the selected data and the data that is not selected. The last selection is a randomly sampled parameter. If you select a fixed parameter, the result of each output is the same, if you leave the default setting, it is not selected, each time you will output a different data.
Percent Sampling:
Row sampling:
"Bi thing" data stream conversion--percent sampling, row sampling