Introduction

Spark SQL provides built-in standard map functions defined in the DataFrame API, which come in handy when you need to perform operations on map (MapType) columns. All these functions accept input as a map column and several other arguments based on the functions.

Spark SQL map functions are grouped as “collection_funcs” in Spark SQL along with several array functions. These map functions are useful when you want to concatenate two or more map columns, convert arrays of StructType entries to a map column, etc.

| Function | Description | | ------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | map_keys | Returns an array containing the keys of the map. | | map_values | Returns an array containing the values of the map. | | map_concat | Merges maps specified in the arguments. | | map_from_entries | Returns a map from the given array of StructType entries. | | map_entries | Returns an array of all StructType in the given map. | | explode() | Creates a new row for every key-value pair in the map by ignoring null & empty values. It creates two new columns, one for key and one for value. | | explode_outer() | Creates a new row for every key-value pair in the map, including null & empty values. It creates two new columns, one for key and one for value. | | posexplode() | Creates a new row for each key-value pair in a map by ignoring null & empty values. It also creates 3 columns: “pos” to hold the position of the map element, and “key” and “value” columns for every row. | | posexplode_outer() | Creates a new row for each key-value pair in a map, including null & empty values. It also creates 3 columns: “pos” to hold the position of the map element, and “key” and “value” columns for every row. | | transform_keys() | Transforms a map by applying functions to every key-value pair and returns a transformed map. | | transform_value() | Transforms a map by applying functions to every key-value pair and returns a transformed map. | | map_zip_with() | Merges two maps into a single map. | | element_at() | Returns a value of a key in a map. | | size() | Returns the length of a map column. |

Example: Using map_keys and map_values

import org.apache.spark.sql.Dataset;
import org.apache.spark.sql.Row;
import org.apache.spark.sql.SparkSession;
import org.apache.spark.sql.functions;
import org.apache.spark.sql.types.DataTypes;
import org.apache.spark.sql.types.StructField;
import org.apache.spark.sql.types.StructType;
import static org.apache.spark.sql.functions.*;

import java.util.Arrays;
import java.util.List;
import java.util.Map;

public class SparkMapFunctions {
    public static void main(String[] args) {
        SparkSession spark = SparkSession.builder()
                .appName("Java Spark SQL Map Functions")
                .config("spark.master", "local")
                .getOrCreate();

        List<Row> data = Arrays.asList(
                RowFactory.create(1, Map.of("key1", "value1", "key2", "value2")),
                RowFactory.create(2, Map.of("key3", "value3", "key4", "value4"))
        );

        StructType schema = new StructType(new StructField[]{
                DataTypes.createStructField("id", DataTypes.IntegerType, false),
                DataTypes.createStructField("map", DataTypes.createMapType(DataTypes.StringType, DataTypes.StringType), false)
        });

        Dataset<Row> df = spark.createDataFrame(data, schema);
        df.select(col("id"), map_keys(col("map")).as("keys"), map_values(col("map")).as("values")).show(false);
    }
}

+---+----------+----------+
|id |keys      |values    |
+---+----------+----------+
|1  |[key1,key2]|[value1,value2]|
|2  |[key3,key4]|[value3,value4]|
+---+----------+----------+

Example: Using map_concat

import org.apache.spark.sql.*;

import java.util.Arrays;
import java.util.List;
import java.util.Map;

import static org.apache.spark.sql.functions.*;

public class SparkMapConcat {
    public static void main(String[] args) {
        SparkSession spark = SparkSession.builder()
                .appName("Java Spark SQL Map Concat")
                .config("spark.master", "local")
                .getOrCreate();

        List<Row> data1 = Arrays.asList(
                RowFactory.create(1, Map.of("key1", "value1")),
                RowFactory.create(2, Map.of("key2", "value2"))
        );

        List<Row> data2 = Arrays.asList(
                RowFactory.create(1, Map.of("key3", "value3")),
                RowFactory.create(2, Map.of("key4", "value4"))
        );

        StructType schema = new StructType(new StructField[]{
                DataTypes.createStructField("id", DataTypes.IntegerType, false),
                DataTypes.createStructField("map", DataTypes.createMapType(DataTypes.StringType, DataTypes.StringType), false)
        });

        Dataset<Row> df1 = spark.createDataFrame(data1, schema);
        Dataset<Row> df2 = spark.createDataFrame(data2, schema);

        Dataset<Row> df = df1.join(df2, "id");
        df.withColumn("map_concat", map_concat(col("map"), df2.col("map"))).show(false);
    }
}

+---+----------+----------+-------------------------+
|id |map1      |map2      |map_concat               |
+---+----------+----------+-------------------------+
|1  |[key1 -> value1]|[key3 -> value3]|[key1 -> value1, key3 -> value3]|
|2  |[key2 -> value2]|[key4 -> value4]|[key2 -> value2, key4 -> value4]|
+---+----------+----------+-------------------------+

Example: Using map_from_entries

import org.apache.spark.sql.*;

import java.util.Arrays;
import java.util.List;

import static org.apache.spark.sql.functions.*;

public class SparkMapFromEntries {
    public static void main(String[] args) {
        SparkSession spark = SparkSession.builder()
                .appName("Java Spark SQL Map From Entries")
                .config("spark.master", "local")
                .getOrCreate();

        List<Row> data = Arrays.asList(
                RowFactory.create(1, new Tuple2[]{new Tuple2<>("key1", "value1"), new Tuple2<>("key2", "value2")}),
                RowFactory.create(2, new Tuple2[]{new Tuple2<>("key3", "value3"), new Tuple2<>("key4", "value4")})
        );

        StructType schema = new StructType(new StructField[]{
                DataTypes.createStructField("id", DataTypes.IntegerType, false),
                DataTypes.createStructField("entries", DataTypes.createArrayType(DataTypes.createStructType(new StructField[]{
                        DataTypes.createStructField("_1", DataTypes.StringType, false),
                        DataTypes.createStructField("_2", DataTypes.StringType, false)
                })), false)
        });

        Dataset<Row> df = spark.createDataFrame(data, schema);
        df.withColumn("map", map_from_entries(col("entries"))).show(false);
    }
}

+---+------------------------+-------------------------+
|id |entries                 |map                      |
+---+------------------------+-------------------------+
|1  |[[key1,value1],[key2,value2]]|[key1 -> value1, key2 -> value2]|
|2  |[[key3,value3],[key4,value4]]|[key3 -> value3, key4 -> value4]|
+---+------------------------+-------------------------+

Example: Using posexplode

import org.apache.spark.sql.*;

import java.util.Arrays;
import java.util.List;
import java.util.Map;

import static org.apache.spark.sql.functions.*;

public class SparkPosExplode {
    public static void main(String[] args) {
        SparkSession spark = SparkSession.builder()
                .appName("Java Spark SQL PosExplode")
                .config("spark.master", "local")
                .getOrCreate();

        List<Row> data = Arrays.asList(
                RowFactory.create(1, Map.of("key1", "value1", "key2", "value2")),
                RowFactory.create(2, Map.of("key3", "value3", "key4", "value4"))
        );

        StructType schema = new StructType(new StructField[]{
                DataTypes.createStructField("id", DataTypes.IntegerType, false),
                DataTypes.createStructField("map", DataTypes.createMapType(DataTypes.StringType, DataTypes.StringType), false)
        });

        Dataset<Row> df = spark.createDataFrame(data, schema);
        df.select(col("id"), posexplode(col("map"))).show(false);
    }
}

+---+---+----+------+
|id |pos|key |value |
+---+---+----+------+
|1  |0  |key1|value1|
|1  |1  |key2|value2|
|2  |0  |key3|value3|
|2  |1  |key4|value4|
+---+---+----+------+