我想从R数据框中的一列生成5个名称的组合，其在不同列中的值加起来等于或小于一定数量

Question

I have a data frame (UFC) with 4 columns.

Column 1 (UFC$Name) is names of UFC fighters fighting this weekend.

Column 2 (UFC$Salary) is how much they "cost" in a fantasy sports contest.

Column 3 (UFC$WinPct) is how likely the fighter is to win the fight.

Column 4 (UFC$FinishPct) is how likely the fighter is to win the fight without it going to a decision.

I'd like to make a data frame that contains all (or more practically the top X number of them, based on the parameter I mention in the next paragraph) the combinations of 5 fighters from column 1, whose column 2 sums add up to $50,000 or less.

Then what I'm really interested in, is the combinations of 5 Fighters whose column 4 sums are highest.

I'm getting pretty good at low level tinkering with data frames but this is a little too advanced for me to wrap my head around how to approach.

Here is about 30% of the dataframe.

              Name Salary WinPct FinishPct
    Keita Nakamura   9100  31.00     15.36
       George Roop   8900  33.00     15.76
   Teruto Ishihara   9000  33.00     17.08
    Naoyuki Kotani   8700  30.50     18.35
     Yusuke Kasuya   8500  29.60     21.16
  Katsunori Kikuno   8800  33.66     21.88

The desired output would look something like this:

Lineup                                                                       
Roy Nelson,Gegard Mousasui,Yusuke Kasuya,George Roop,Diego Brandao      
SalarySum
47900     
FinishPctSum     
148.99 

And it would return the top X number of those outputs, ranked by highest FinishPctSum

Answer 1

Well this won't be terribly fast but it's an idea ...

## make a list of all combinations of 5 of Name, Salary, and FinishPct
xx <- with(df, lapply(list(as.character(Name), Salary, FinishPct), combn, 5))
## convert the names to a string, 
## find the column sums of the others,
## set the names
yy <- setNames(
    lapply(xx, function(x) {
        if(typeof(x) == "character") apply(x, 2, toString) else colSums(x)
    }),
    names(df)[c(1, 2, 4)]
)
## coerce to data.frame
newdf <- as.data.frame(yy)

which results in

#                                                                              Names Salary FinishPct
# 1      Keita Nakamura, George Roop, Teruto Ishihara, Naoyuki Kotani, Yusuke Kasuya  44200     87.71
# 2   Keita Nakamura, George Roop, Teruto Ishihara, Naoyuki Kotani, Katsunori Kikuno  44500     88.43
# 3    Keita Nakamura, George Roop, Teruto Ishihara, Yusuke Kasuya, Katsunori Kikuno  44300     91.24
# 4     Keita Nakamura, George Roop, Naoyuki Kotani, Yusuke Kasuya, Katsunori Kikuno  44000     92.51
# 5 Keita Nakamura, Teruto Ishihara, Naoyuki Kotani, Yusuke Kasuya, Katsunori Kikuno  44100     93.83
# 6    George Roop, Teruto Ishihara, Naoyuki Kotani, Yusuke Kasuya, Katsunori Kikuno  43900     94.23

No check has been performed to determine whether the salaries are less than 50k. It just gives all the combinations of 5 fighters with their respective sums. You can subset to find those salaries less than 50k with

newdf[newdf$Salary <= 5e4, ]

Note that 5e4 is shorthand/scientific notation for 50,000.

Data:

df <- structure(list(Name = structure(c(3L, 1L, 5L, 4L, 6L, 2L), .Label = c("George Roop", 
"Katsunori Kikuno", "Keita Nakamura", "Naoyuki Kotani", "Teruto Ishihara", 
"Yusuke Kasuya"), class = "factor"), Salary = c(9100L, 8900L, 
9000L, 8700L, 8500L, 8800L), WinPct = c(31, 33, 33, 30.5, 29.6, 
33.66), FinishPct = c(15.36, 15.76, 17.08, 18.35, 21.16, 21.88
)), .Names = c("Name", "Salary", "WinPct", "FinishPct"), class = "data.frame", row.names = c(NA, 
-6L))