寻找一种比在R中循环更有效的方法来创建数据帧

Question

I am trying to create a table/dataframe based on the AcceptanceSampling library as follows:

library(AcceptanceSampling)
df<-NULL
for (aql in c(0.01,0.05)){
for (prp in c(0.95)) {
for (def in c(0.06,0.1,0.15)){
for (crp in c(0.05,0.08,0.10)){
df<-as.data.frame(rbind(df,c(aql,prp,def,crp,
                         find.plan(PRP=c(aql,prp),CRP=c(def,crp))$n,
                         find.plan(PRP=c(aql,prp),CRP=c(def,crp))$c
                         )))
}}}}

names(df)<-c("aql","prp","def","crp","n","Ac")

this gives me:

    aql  prp  def  crp    n  Ac
1  0.01 0.95 0.06 0.05  127   3
2  0.01 0.95 0.06 0.08  116   3
3  0.01 0.95 0.06 0.10  110   3
4  0.01 0.95 0.10 0.05   61   2
5  0.01 0.95 0.10 0.08   55   2
6  0.01 0.95 0.10 0.10   52   2
7  0.01 0.95 0.15 0.05   30   1
8  0.01 0.95 0.15 0.08   27   1
9  0.01 0.95 0.15 0.10   25   1
10 0.05 0.95 0.06 0.05 5626 308
11 0.05 0.95 0.06 0.08 4826 266
12 0.05 0.95 0.06 0.10 4445 246
13 0.05 0.95 0.10 0.05  298  21
14 0.05 0.95 0.10 0.08  251  18
15 0.05 0.95 0.10 0.10  233  17
16 0.05 0.95 0.15 0.05   93   8
17 0.05 0.95 0.15 0.08   79   7
18 0.05 0.95 0.15 0.10   77   7

Can someone point to a more efficient way to build this ? Preferably without the loops and without having to call find.plan() twice for each row ?

Thanks in advance Pete

Answer 1

You can use expand.grid like this :

dat <- expand.grid(aql = c(0.01,0.05),prp = c(0.95),
            def = c(0.06,0.1,0.15), crp = c(0.05,0.08,0.10))

Then using data.table for each syntax sugar:

library(data.table)
DT <- as.data.table(dat)
DT[, c('n','Ac') := list(find.plan(PRP=c(aql,prp),CRP=c(def,crp))$n,
                         find.plan(PRP=c(aql,prp),CRP=c(def,crp))$c),
                           by = 1:nrow(DT)]

    aql  prp  def  crp    n  Ac
 1: 0.01 0.95 0.06 0.05  127   3
 2: 0.05 0.95 0.06 0.05 5626 308
 3: 0.01 0.95 0.10 0.05   61   2
 4: 0.05 0.95 0.10 0.05  298  21
 5: 0.01 0.95 0.15 0.05   30   1
 6: 0.05 0.95 0.15 0.05   93   8
 7: 0.01 0.95 0.06 0.08  116   3
 8: 0.05 0.95 0.06 0.08 4826 266
 9: 0.01 0.95 0.10 0.08   55   2
10: 0.05 0.95 0.10 0.08  251  18
11: 0.01 0.95 0.15 0.08   27   1
12: 0.05 0.95 0.15 0.08   79   7
13: 0.01 0.95 0.06 0.10  110   3
14: 0.05 0.95 0.06 0.10 4445 246
15: 0.01 0.95 0.10 0.10   52   2
16: 0.05 0.95 0.10 0.10  233  17
17: 0.01 0.95 0.15 0.10   25   1
18: 0.05 0.95 0.15 0.10   77   7

EDIT using base functions, the idea is to vectorize find.plan . Here I am using mapply like this:

 cbind(dat,with(dat,t(mapply(function(x,y,z,t)
                 find.plan(c(x,y),c(z,t)),aql,prp,def,crp))))

  aql  prp  def  crp    n   c   r
1  0.01 0.95 0.06 0.05  127   3   4
2  0.05 0.95 0.06 0.05 5626 308 309
3  0.01 0.95 0.10 0.05   61   2   3
4  0.05 0.95 0.10 0.05  298  21  22
5  0.01 0.95 0.15 0.05   30   1   2
6  0.05 0.95 0.15 0.05   93   8   9
7  0.01 0.95 0.06 0.08  116   3   4
8  0.05 0.95 0.06 0.08 4826 266 267
9  0.01 0.95 0.10 0.08   55   2   3
10 0.05 0.95 0.10 0.08  251  18  19
11 0.01 0.95 0.15 0.08   27   1   2
12 0.05 0.95 0.15 0.08   79   7   8
13 0.01 0.95 0.06 0.10  110   3   4
14 0.05 0.95 0.06 0.10 4445 246 247
15 0.01 0.95 0.10 0.10   52   2   3
16 0.05 0.95 0.10 0.10  233  17  18
17 0.01 0.95 0.15 0.10   25   1   2
18 0.05 0.95 0.15 0.10   77   7   8

Answer 2

An alternative answer using base R functions:

install.packages("AcceptanceSampling")
library(AcceptanceSampling)

df <- expand.grid(
  aql = c(0.01,0.05),
  prp = c(0.95),
  def = c(0.06,0.1,0.15),
  crp = c(0.05,0.08,0.10)
)

findpl <- do.call(
  rbind,
  by(df,df,function(x) {
    i <- find.plan(c(x$aql,x$prp),c(x$def,x$crp))
    c(n=i$n,Ac=i$c)
    }
  )
)

result <- data.frame(df,findpl)

head(result)

   aql  prp  def  crp    n  Ac
1 0.01 0.95 0.06 0.05  127   3
2 0.05 0.95 0.06 0.05 5626 308
3 0.01 0.95 0.10 0.05   61   2
4 0.05 0.95 0.10 0.05  298  21
5 0.01 0.95 0.15 0.05   30   1
6 0.05 0.95 0.15 0.05   93   8