两个矩阵/数据帧的元素级联

Question

Say I have a df:

df=data.frame(
    A=rep(c('a1','a1','a2','a2'),10),
    B=rep(c('b1','b2','b2','b2'),10)
   )

and I want a table like this:

What I can do is using table , addMargins and prop.table to get two separate matrices, but I don't know if there has any efficient way to concatenate 2 matrices into the format I provide.

>df%>%table()%>%addmargins()

     B
A     b1 b2 Sum
  a1  10 10  20
  a2   0 20  20
  Sum 10 30  40

>df%>%table()%>%prop.table(margin=2)

    B
A         b1      b2
  a1 1.00000 0.33333
  a2 0.00000 0.66667

Answer 1

We can use sprintf as shown:

tab.sum <- table(df)
tab.final <-  addmargins(tab.sum)
tab.prop <- prop.table(tab.sum, margin=2)
tab.final[1:2, 1:2] <- sprintf("%d (%.2f%%)", tab.sum, tab.prop)
tab.final

giving:

     B
A     b1         b2         Sum
  a1  10 (1.00%) 10 (0.33%) 20 
  a2  0 (0.00%)  20 (0.67%) 20 
  Sum 10         30         40 

however, you might want to try these packages:

1) Try CrossTable in gmodels. It has many arguments that can be used to customize the output.

library(gmodels)
with(df, CrossTable(A, B))

giving:

   Cell Contents
|-------------------------|
|                       N |
| Chi-square contribution |
|           N / Row Total |
|           N / Col Total |
|         N / Table Total |
|-------------------------|


Total Observations in Table:  40 


             | B 
           A |        b1 |        b2 | Row Total | 
-------------|-----------|-----------|-----------|
          a1 |        10 |        10 |        20 | 
             |     5.000 |     1.667 |           | 
             |     0.500 |     0.500 |     0.500 | 
             |     1.000 |     0.333 |           | 
             |     0.250 |     0.250 |           | 
-------------|-----------|-----------|-----------|
          a2 |         0 |        20 |        20 | 
             |     5.000 |     1.667 |           | 
             |     0.000 |     1.000 |     0.500 | 
             |     0.000 |     0.667 |           | 
             |     0.000 |     0.500 |           | 
-------------|-----------|-----------|-----------|
Column Total |        10 |        30 |        40 | 
             |     0.250 |     0.750 |           | 
-------------|-----------|-----------|-----------|

2) Also CrossTable in descr.

library(descr)

with(df, CrossTable(A, B))

   Cell Contents 
|-------------------------|
|                       N | 
| Chi-square contribution | 
|           N / Row Total | 
|           N / Col Total | 
|         N / Table Total | 
|-------------------------|

==============================
         B
A           b1      b2   Total
------------------------------
a1          10      10      20
         5.000   1.667        
           0.5     0.5     0.5
         1.000   0.333        
          0.25    0.25        
------------------------------
a2           0      20      20
         5.000   1.667        
           0.0     1.0     0.5
         0.000   0.667        
          0.00    0.50        
------------------------------
Total       10      30      40
          0.25    0.75        
==============================

3) The tables package has a formula interface. It also supports latex output.

tabular((A + 1)
*((n=1) + Percent()
+ (RowPct=Percent("row"))
+ (ColPct=Percent("col")))
~ (B + 1)
*Format(digits=1), data=df ) 

giving:

             B          
 A           b1  b2  All
 a1  n        10  10  20
     Percent  25  25  50
     RowPct   50  50 100
     ColPct  100  33  50
 a2  n         0  20  20
     Percent   0  50  50
     RowPct    0 100 100
     ColPct    0  67  50
 All n        10  30  40
     Percent  25  75 100
     RowPct   25  75 100
     ColPct  100 100 100

4) janitor

The janitor package supports a pipeline specification and supports piping its output to kable.

library(janitor)

df %>%
  tabyl(A, B) %>%
  adorn_totals(c("row", "col")) %>%
  adorn_percentages("all") %>%
  adorn_pct_formatting(digits = 1) %>%
  adorn_ns %>%
  adorn_title

giving:

                B                       
     A         b1         b2       Total
    a1 25.0% (10) 25.0% (10)  50.0% (20)
    a2  0.0%  (0) 50.0% (20)  50.0% (20)
 Total 25.0% (10) 75.0% (30) 100.0% (40)