Question

说我有以下Frame，

type Person = 
   { Name:string; Age:int; Comp1:float; Comp2:float }

let peopleRecds = 
    [ { Name = "Joe"; Age = 51; Comp1=12.1; Comp2 =20.3 }
      { Name = "Tomas"; Age = 28; Comp1=1.1; Comp2 =29.3 }
      { Name = "Eve"; Age = 2; Comp1=2.1; Comp2 =40.3 }
      { Name = "Suzanne"; Age = 15; Comp1=12.4; Comp2 =26.3} ]
let peopleList = Frame.ofRecords peopleRecds

我想要做的是将Comp1和Comp2列合并到peopleList的新列中，n的{{1}}尚未确定目前，所以我不能只知道总和两列，可能有Comp(n)，Comp3，因此必须以正则表达为基础，类似于密钥导致Comp4。

我应该做的就是每行Comp

mapRowValues

但是我不知道如何在这里的行级操作。

Answer 1

如果计划增加＆＃39; Comp＆＃39; ，把它们放在一个数组中可能会更好：

type nPerson =  { Name:string; Age:int; Comp:float[] }

let npeopleRecds = 
    [ { Name = "Joe"; Age = 51; Comp = [| 12.1; 20.3 |] }
      { Name = "Tomas"; Age = 28; Comp = [| 1.1; 29.3  |] }
      { Name = "Eve"; Age = 2; Comp = [| 2.1; 40.3  |] }
      { Name = "Suzanne"; Age = 15; Comp = [|12.4; 26.3 |] } ]

然后，您可以轻松添加总和为

的列

let npeopleList = Frame.ofRecords npeopleRecds

npeopleList.Format() |> printfn "%s"

let sumseries = npeopleList.GetColumn<float []>("Comp") |> Series.mapValues(fun x -> x |> Array.sum)

npeopleList?TotalComp <- sumseries

npeopleList.Format() |> printfn "%s"

打印：

     Name    Age Comp
0 -> Joe     51  System.Double[]
1 -> Tomas   28  System.Double[]
2 -> Eve     2   System.Double[]
3 -> Suzanne 15  System.Double[]

     Name    Age Comp            TotalComp
0 -> Joe     51  System.Double[] 32,4
1 -> Tomas   28  System.Double[] 30,4
2 -> Eve     2   System.Double[] 42,4
3 -> Suzanne 15  System.Double[] 38,7

修改

如果更改人员记录中的字段是不可接受的 - 您可以使用过滤器：

let allSum = peopleList.Columns |> Series.filter(fun k _ -> k.StartsWith("Comp")) |> Frame.ofColumns |> Frame.rows |> Series.mapValues(Series.foldValues(fun acc v -> acc + (v :?> float)) 0.0) peopleList?TotalComp <- allSum peopleList.Format() |> printfn "%s"

打印：

Name Age Comp1 Comp2 TotalComp 0 -> Joe 51 12,1 20,3 32,4 1 -> Tomas 28 1,1 29,3 30,4 2 -> Eve 2 2,1 40,3 42,4 3 -> Suzanne 15 12,4 26,3 38,7

f＃deedle基于列键聚合行值

1 个答案: