df = spark.createDataFrame([(‘LC7-H6116BCF-R-GL-201116V750Fans‘, ‘张三‘, 88), (‘语文‘, ‘张三‘, 92), (‘英语‘, ‘张三‘, 77),
                            (‘数学‘, ‘王五‘, 65), (‘语文‘, ‘王五‘, 87), (‘英语‘, ‘王五‘, 90),
                            (‘数学‘, ‘李雷‘, 67), (‘语文‘, ‘李雷‘, 33), (‘英语‘, ‘李雷‘, 24),
                            (‘数学‘, ‘宫九‘, 77), (‘语文‘, ‘宫九‘, 87)
                            ], [‘subject‘, ‘name‘, ‘score‘])

df.createOrReplaceTempView(‘t1‘)
spark.sql(
    "select subject,collect_set(Array(name,score)) as c"
    " from t1"
    " group by subject"
).show()