这是我将采取的方法：
构建一个dataframe，它为您想要表示的每一分钟都有一行（我使用spark.range为我需要的每一分钟投影一行）
对于范围内的每个id，在开始日期前加一分钟
使用左外连接将日期连接到原始Dataframe，这样就不会丢失任何行
然后使用last来填充任何空白-注意，如果以null开头，则newvalue将为null，直到获得一些数据

using System;
using System.Collections.Generic;
using Microsoft.Spark.Sql;
using Microsoft.Spark.Sql.Expressions;
using Microsoft.Spark.Sql.Types;

namespace StackOverflow
{
    class Program
    {
        static void Main(string[] args)
        {
            var spark = SparkSession.Builder().GetOrCreate();

            //Sample data set - we will fill in the missing minutes
            var df = spark.CreateDataFrame(new List<GenericRow>()
            {
                new GenericRow(new object[] {2021, 3, 4, 8, 3, 87, "Type1", 380.5}),
                new GenericRow(new object[] {2021, 3, 4, 8, 10, null, null, null}),
                new GenericRow(new object[] {2021, 3, 4, 8, 20, null, null, null}),
                new GenericRow(new object[] {2021, 3, 4, 8, 25, null, null, null}),
                new GenericRow(new object[] {2021, 3, 4, 8, 35, 87, "Type1", 0.0}),
                new GenericRow(new object[] {2021, 3, 4, 8, 45, 87, "Type1", 0.0})
            }, new StructType(new List<StructField>()
            {
                new StructField("Year", new IntegerType()),
                new StructField("Month", new IntegerType()),
                new StructField("Day", new IntegerType()),
                new StructField("Hour", new IntegerType()),
                new StructField("Minute", new IntegerType()),
                new StructField("ID", new IntegerType()),
                new StructField("Type", new StringType()),
                new StructField("Value", new DoubleType()),
            }));

            //start and end time
            var reportFrom = new DateTime(2021, 3, 4, 7, 0, 0);
            var reportTo = new DateTime(2021, 3, 4, 9, 0, 0);

            //convert start time to unix epoch as we can't pass a DateTime to spark (yet!)
            var unixFromTime = (reportFrom - new DateTime(1970, 1, 1, 0, 0, 0, 0)).TotalSeconds;

            //how many total rows do we need?
            var minutesToCreate = reportTo.Subtract(reportFrom).TotalMinutes;

            //create a dataframe with 1 row for every minute we need
            var everyMinute = spark.Range((long) minutesToCreate);

            //Add the reportFrom unix epoch
            everyMinute = everyMinute.WithColumn("BaseTime", Functions.Lit(unixFromTime));

            //add to the unix epoch, the Id (incrementing number) multiplied by 60 - if we didn't mul(60) it would add seconds and not minutes
            everyMinute = everyMinute.WithColumn("Time",
                Functions.Lit(unixFromTime)
                    .Plus(Functions.Col("Id").Cast("Int").Multiply(Functions.Lit(60))));

            //convert the unix epoch to an actual timestamp and drop all the intermediate columns
            everyMinute = everyMinute.WithColumn("Date",
                Functions.ToTimestamp(Functions.FromUnixTime(Functions.Col("Time")))).Select("Date");

            //convert timestamp into individual columns

            everyMinute = everyMinute.WithColumn("Year", Functions.Year(Functions.Col("Date")));
            everyMinute = everyMinute.WithColumn("Month", Functions.Month(Functions.Col("Date")));
            everyMinute = everyMinute.WithColumn("Day", Functions.DayOfMonth(Functions.Col("Date")));
            everyMinute = everyMinute.WithColumn("Hour", Functions.Hour(Functions.Col("Date")));
            everyMinute = everyMinute.WithColumn("Minute", Functions.Minute(Functions.Col("Date")));

            //join both data frames so...
            var dfAllData = everyMinute.Join(df, new List<string>() {"Year", "Month", "Day", "Hour", "Minute"}, "left_outer");

            //add in data using Last
            var window = Window.OrderBy("Year", "Month", "Day", "Hour", "Minute");
            var filledDataFrame = dfAllData.WithColumn("newValue",
                Functions.When(dfAllData["Value"].IsNull(),
                        Functions.Last(dfAllData["Value"], true).Over(window))
                    .Otherwise(dfAllData["Value"]));

            filledDataFrame.Show(1000, 10000);
        }
    }
}

预计起飞时间