Question

I am attempting to create a data set from an original dataframe (in either R or Excel VBA code), and I am trying to generate another column. Here is a high-level situation:

dfr <- data.frame(
    grp   = rep(c("X", "Y"), each = 4),
    id    = c("A", "B", "C", "D", "E", "A", "B", "F"),
    value = c(3, 7, 2, 4, 8, 9, 11, 2)
) 

Of the second column, B is the “leader” of both group “X” and “Y” as it has the largest numbers. Thus, I need to pair all the observations other observations in the relevant groups (X, Y) to these leaders. For example, a sample output is below of what I need:

X  B A 3
X  B C 2
X  B D 4
Y  B E 8
Y  B A 9
Y  B F 2

The number on the furthest column is the respective number of the datapoint previously found.

So, I need help partitioning the data between X and Y (for countless string groups that exist) and afterwards, sorting it in the way I need and having that column produced, either in R-code or VBA for excel (the data is in CSV format)

**Disclaimer: If it isn’t obvious, my use of R is very limited – I used it for 4 months in an Applied Econometrics course, and now am finding myself in need of it again (9months later) so please excuse me if I seem like a novice…I can run regressions just fine though 🙂

*UPDATE
Following Henry’s code, this is where I am now.

data <- read.csv(file = "sort.csv", h=T)
attach(data)
sorted <- data[order(data$membernumber, -data$dailycirc),]
top <- function(df){ return(df[1,])}
moded <- unsplit(lapply(split(sorted, sorted$membernumber), top), unique(sorted$membernumber))[1:2]
names(moded) <- c("membernumber", "cnty")
merged <- merge(moded, data, by="membernumber")
merged[merged$cnty != merged$cnty, ]
summary(merged)

This has actually now given me some output. But I’m not seeing the sort, just a summary statistic of things like the mean/max. How do I actually export this into a CSV or spreadsheet so I can look at it like a table?

Thank you SO much for your help.

Answer 1

Since only R people seem to have waken up I felt obliged to post an example in VBA. I assume that you have opened your CSV in Excel with information in the first 3 columns, no title line (data starts on row 1). You copy that
data to a new spreadsheet, in a tab named “Sheet1”. You insert the code below in a new module and run it, and it will output the result in columns 5 to 8 of the same sheet. For 10,000 lines input, it runs in less than 0.1 sec on my machine.

Option Explicit

Public Sub doIt()

    Dim data As Variant
    Dim result As Variant
    Dim i As Long
    Dim j As Long
    Dim dict1 As Variant
    Dim dict2 As Variant

    Set dict1 = CreateObject("Scripting.Dictionary")
    Set dict2 = CreateObject("Scripting.Dictionary")
    data = Sheets("Sheet1").UsedRange

    For i = LBound(data, 1) To UBound(data, 1)
        If dict1.exists(data(i, 1)) Then
            If dict2(data(i, 1)) < data(i, 3) Then
                dict1(data(i, 1)) = data(i, 2)
                dict2(data(i, 1)) = data(i, 3)
            End If
        Else
            dict1(data(i, 1)) = data(i, 2)
            dict2(data(i, 1)) = data(i, 3)
        End If
    Next i

    ReDim result(LBound(data, 1) To UBound(data, 1) - dict1.Count, 1 To 4) As Variant

    j = 1
    For i = LBound(data, 1) To UBound(data, 1)
        If data(i, 2) <> dict1(data(i, 1)) Then
            result(j, 1) = data(i, 1)
            result(j, 2) = dict1(data(i, 1))
            result(j, 3) = data(i, 2)
            result(j, 4) = data(i, 3)
            j = j + 1
        End If
    Next i

    With Sheets("Sheet1")
        .Cells(1, 5).Resize(UBound(result, 1), UBound(result, 2)) = result
    End With

End Sub