小编fac*_*eID的帖子

假设这个代码：

public static Dataset<Row> getData(SparkSession sparkSession,
                                             StructType schema, String delimiter, String pathToData) {

    final Dataset<Row> dataset = sparkSession
            .read()
            .option("delimiter", "\\t")
            .csv(pathToData);

    StructType nSchema= newSchema(schema, schema.size(), dataset.columns().length);
 ...
}

在将变量提供给 newSchema 方法之前声明变量并使它们成为 final 是最佳实践吗？

public static Dataset<Row> getData(SparkSession sparkSession,
                                             StructType schema, String delimiter, String pathToData) {

    final Dataset<Row> dataset = sparkSession
            .read()
            .option("delimiter", "\\t")
            .csv(pathToData);

    final int dataSize = dataset.columns().length;
    final int schemaSize = schema.size();

    StructType nSchema = newSchema(schema, schemaSize, dataSize);
 ...
}

谢谢