如何使用Java读取Spark中的xls和xlsx文件？

Question

I want to read xls and xlsx (MS Excel) files row by row in spark like we do it for text files OR any how?

I want to use spark to increase performance for reading a large xls file say 1 GB, that's why I need spark to read the file in parts like we do for text files.

How can I read data from excel files in spark whether it is line by line or not?

I just want to read entries in the xls file anyhow using spark.

Please suggest.

Thanks!!!

Answer 1

You can not do that with spark . It is not meant for it. Use someother library eg Apache POI to read excel and then feed that data to spark as text.

Answer 2

Though the question is bit old, i am still answering it. May be it will be useful to someone else. The answer is yes you can do it with apache spark 2.x. Let say you want to convert a xls with 3 columns to Dataset.

  class Bean {
     private String col1;
     private String col2;   
     private Timestamp col3;
}

StructType structType= new StructType(new StructField[] {
                new StructField("col1", DataTypes.StringType, true, Metadata.empty()),
                new StructField("col2", DataTypes.StringType, true, Metadata.empty()),
                new StructField("col3", DataTypes.TimestampType, true, Metadata.empty())
        });

Dataset<Bean> ds = sparkSession.read().
                schema(structType).
                format("com.crealytics.spark.excel").
                option("useHeader", true). // If the xls file has headers
                option("timestampFormat", "yyyy-MM-dd HH:mm:ss"). // If you want to convert timestamp to a specific format
                option("treatEmptyValuesAsNulls", "false").
                option("inferSchema", "false").
                option("addColorColumns", "false").
                load("/home/user/test/sample.xls"). //path to xls or xlsx
                as(Encoders.bean(Bean.class)); // Bean in which you want to convert the data, you can remove this line if Dataset<Row> is just fine for you

Answer 3

You can try the HadoopOffice library to read/write Excel files with Spark ( https://github.com/ZuInnoTe/hadoopoffice/wiki ). It supports encrypted Excel, linked workbooks, filtering by metadata ...

Answer 4

Here is how I done.

In maven add dependencies

<dependencies>
    <dependency>
        <groupId>org.apache.spark</groupId>
        <artifactId>spark-core_2.11</artifactId>
        <version>2.4.2</version>
    </dependency>
    <dependency>
        <groupId>org.apache.spark</groupId>
        <artifactId>spark-sql_2.11</artifactId>
        <version>2.4.2</version>
    </dependency>
    <dependency>
        <groupId>com.crealytics</groupId>
        <artifactId>spark-excel_2.11</artifactId>
        <version>0.11.1</version>
    </dependency>
</dependencies>

My main class

import org.apache.log4j.Level;
import org.apache.log4j.Logger;
import org.apache.spark.sql.Dataset;
import org.apache.spark.sql.Row;
import org.apache.spark.sql.SparkSession;

public class ReadExcelSheets {

    public static void main(String[] args) {
        //skip logging extras
        Logger.getLogger("org").setLevel(Level.ERROR);

       //build session
        SparkSession spark = SparkSession
                .builder()
                .appName("Java Spark SQL Example")
                .config("spark.master", "local")
                .getOrCreate();

        //read excel - change file name
        Dataset<Row> df = spark.read()
                .format("com.crealytics.spark.excel")
                .option("useHeader", "true")
                //.option("dataAddress", "'Sheet1'!A1:M1470") // optional when you want to read sheets where A1 first top cell and M1470 us very bottom left of sheet.
                .load("datasets/test1.xlsx");
        //show your data
        df.show();
    }
}