刘春霞：大数据成功关键.pptVIP

* Here are the properties for the two BDFS stages on the left of the previous diagram, the BigInsights data sources. You can see how we easily point to the appropriate BigInsights servers, specify the files we need, the data format we expect to receive and we’re up and running. You can see two read methods are being used in these examples. On the left we use the single file or specific file method, on the right we use the file pattern method. We can also specify how many readers we need. This is the parallelism we can provide by scaling the number of BDFS operators we trigger for a particul