如何查找数据帧列中值的第一个多次出现之间的最小和最大中间项数

Question

I have a factor vector called Categories with 47 levels

Categories = as.factor(sort(make.unique(rep(letters, length.out = 47), sep='')))

[1] a  a1 b  b1 c  c1 d  d1 e  e1 f  f1 g  g1 h  h1 i  i1 j  j1 k  k1 l  l1 m  m1 n  n1 o  o1 p  p1 q  q1 r  r1 s  s1 t 
[40] t1 u  u1 v  w  x  y  z 
47 Levels: a a1 b b1 c c1 d d1 e e1 f f1 g g1 h h1 i i1 j j1 k k1 l l1 m m1 n n1 o o1 p p1 q q1 r r1 s s1 t t1 u u1 ... z

I have another vector called cat with 9 of those levels

cat = Categories[c(7,42,43,24,45,26,35,6,15)]
[1] d  u1 v  l1 x  m1 r  c1 h 
47 Levels: a a1 b b1 c c1 d d1 e e1 f f1 g g1 h h1 i i1 j j1 k k1 l l1 m m1 n n1 o o1 p p1 q q1 r r1 s s1 t t1 u u1 ... z

I also have a dataframe My_Data with 36 rows. One of the columns in the dataframe has multiple occurences of the values from cat

My_Data = data.frame(name = make.unique(rep(c(1:10,LETTERS), length.out = 36), sep=''), cat = sample(rep(cat,4),36,replace = FALSE), position = 0)

    name cat position
1     1   v        0
2     2   r        0
3     3   h        0
4     4  m1        0
5     5   h        0
6     6  u1        0
7     7  l1        0
8     8   h        0
9     9  u1        0
10   10   x        0
11    A   x        0
12    B   v        0
13    C   d        0
14    D  c1        0
15    E   r        0
16    F   v        0
17    G  l1        0
18    H   d        0
19    I  l1        0
20    J  c1        0
21    K  u1        0
22    L   x        0
23    M   v        0
24    N   d        0
25    O  l1        0
26    P  m1        0
27    Q   r        0
28    R  m1        0
29    S   h        0
30    T  m1        0
31    U  c1        0
32    V   d        0
33    W   r        0
34    X   x        0
35    Y  c1        0
36    Z  u1        0

Using the code below, I can populate the position column given above to reflect the number of occurence of the value in the cat column:

transform(My_Data, position = ave(as.character(cat), cat, FUN = seq_along))

The first 15 rows of the dataframe My_Data would look like:

    name cat position
1     1   v        1
2     2   r        1
3     3   h        1
4     4  m1        1
5     5   h        2
6     6  u1        1
7     7  l1        1
8     8   h        3
9     9  u1        2
10   10   x        1
11    A   x        2
12    B   v        2
13    C   d        1
14    D  c1        1
15    E   r        2

Now I want to calculate the min. and max. number of intervening items between any 2 consecutive occurences of the same value of the cat column.

How can I do this?

Answer 1

If I understand your question, here's one option:

library(tidyverse)

# Data
Categories = as.factor(sort(make.unique(rep(letters, length.out = 47), sep='')))  
cat = Categories[c(7,42,43,24,45,26,35,6,15)]
# Set a seed for reproducibility
set.seed(5)
My_Data = data.frame(name = make.unique(rep(c(1:10,LETTERS), length.out = 36), sep=''), 
                     cat = sample(rep(cat,4),36,replace = FALSE), 
                     position = 0)

The code below summarises to give the minimum and maximum number of intervening rows for each level of cat .

# Summarise to give min and max number rows between each occurrence
My_Data %>%
  mutate(row=1:n()) %>% 
  group_by(cat) %>% 
  summarise(min.diff=min(diff(row)-1, na.rm=TRUE),
            max.diff=max(diff(row)-1, na.rm=TRUE))

  cat min.diff max.diff <fctr> <dbl> <dbl> 1 c1 4 6 2 d 1 16 3 h 1 16 4 l1 0 13 5 m1 0 12 6 r 5 15 7 u1 2 7 8 v 1 16 9 x 6 12 

If you want to mark the number of intervening rows in the original data frame: The code below adds a column to the original data frame to give the number of intervening rows since the last occurrence of a given level of cat .

# Add column with intervening number of rows between each occurrence in cat
My_Data %>%
  mutate(row=1:n()) %>% 
  group_by(cat) %>% 
  mutate(diff=c(NA,diff(row)-1)) %>%
  select(-row)

  name cat position diff <fctr> <fctr> <dbl> <dbl> 1 1 c1 0 NA 2 2 m1 0 NA 3 3 x 0 NA 4 4 d 0 NA 5 5 l1 0 NA 6 6 l1 0 0 7 7 r 0 NA 8 8 c1 0 6 9 9 h 0 NA 10 10 v 0 NA 

Answer 2

Here is a tidy solution using lag() :

library(tidyverse)

# create data frame
set.seed(1)
Categories <- as.factor(sort(make.unique(rep(letters, length.out = 47), sep='')))
cat <- Categories[c(7,42,43,24,45,26,35,6,15)]
My_Data <- data.frame(
  name = make.unique(rep(c(1:10,LETTERS), length.out = 36), sep=''), 
  cat = sample(rep(cat,4),36,replace = FALSE), 
  position = 0
)

# solution
My_Data %>%
  mutate(row = 1:n()) %>%
  group_by(cat) %>%
  mutate(inter = row - lag(row) - 1) %>%
  summarize(min_inter = min(inter, na.rm = T), max_inter = max(inter, na.rm = T))

Result:

# A tibble: 9 x 3
     cat min_inter max_inter
  <fctr>     <dbl>     <dbl>
1     c1         0        10
2      d         4        11
3      h         0         8
4     l1         0         6
5     m1         1         3
6      r         0        16
7     u1         2         5
8      v         1        23
9      x         6        15