根据列数据生成组列

Question

I am new to R, I am trying to introduce a group column based on the data in a column.

Example of the data.frame

1           11.3178501            4    9        11.618880
2           10.3969713           20    8        11.047486
8            9.5067421           14    7        10.079806
6            6.6135932            6    6         7.002669
4            5.4157174            2    5         5.566232
17           3.8860793            5    4         4.235564
16           3.8183699           15    3         4.406416
11           1.2574765           18    2         1.885113
15           0.7084411            7    1         1.130990

First column is the index introduced by the R but I sorted so the order is different, what I am trying to do is introduce a column that defines the bracket that each row is belong to based on the last column value. so if last column value is between 0-5 => 1, 5-0 => 2 etc then we add a new column at the end group -> 1,2,3...

16           3.8183699           15    3         4.406416    1
11           1.2574765           18    2         1.885113    2
15           0.7084411            7    1         1.130990    2

I tried the following dataFrame$column4 < 5 but this gave me a boolean value so I thought I'll multiply that by 1 then i got the following

0 0 0 0 0 1 1 1 1

I am not sure if I am on the right track.

Answer 1

Even given your comment, I would still suggest cut . It is in base R and usually not considered a fancy function.

df <- structure(list(V1 = c(1L, 2L, 8L, 6L, 4L, 17L, 16L, 11L, 15L), 
    V2 = c(11.3178501, 10.3969713, 9.5067421, 6.6135932, 5.4157174, 
    3.8860793, 3.8183699, 1.2574765, 0.7084411), V3 = c(4L, 20L, 
    14L, 6L, 2L, 5L, 15L, 18L, 7L), V4 = c(9L, 8L, 7L, 6L, 5L, 
    4L, 3L, 2L, 1L), V5 = c(11.61888, 11.047486, 10.079806, 7.002669, 
    5.566232, 4.235564, 4.406416, 1.885113, 1.13099)), .Names = c("V1", 
"V2", "V3", "V4", "V5"), class = "data.frame", row.names = c(NA, 
-9L))

df$groups <- cut(df$V5, seq(0,15, by=5))

> df
  V1         V2 V3 V4        V5  groups
1  1 11.3178501  4  9 11.618880 (10,15]
2  2 10.3969713 20  8 11.047486 (10,15]
3  8  9.5067421 14  7 10.079806 (10,15]
4  6  6.6135932  6  6  7.002669  (5,10]
5  4  5.4157174  2  5  5.566232  (5,10]
6 17  3.8860793  5  4  4.235564   (0,5]
7 16  3.8183699 15  3  4.406416   (0,5]
8 11  1.2574765 18  2  1.885113   (0,5]
9 15  0.7084411  7  1  1.130990   (0,5]
> 

Finally, if integers are what you want, you can coerce the groups to integers using factor .

df$groups <- as.integer(df$groups)

> as.integer(df$groups)
[1] 3 3 3 2 2 1 1 1 1

Answer 2

Justin's answer is great; yet if you want to implement dumber cut on your own, you can do this this way. First, you define a vector with your thresholds, like thre<-c(0,5,10,15) , then do an outer comparison of your values and those thresholds with greater-than operator and sum the rows of such created matrix like this:

rowSums(outer(values,thre,'>'))

And voila, all values in (0,5] are now 1, (5,10] are 2, etc.
Wrapped in a function, it could look like this:

ultraDumbCut<-function(v,thre) rowSums(outer(v,thre,'>'))

Made a bit more intelligent, like this:

dumbCut<-function(v,jump=5,thre=seq(0,max(v),by=jump)) rowSums(outer(v,thre,'>'))

so that dumbCut(1:7) is 1 1 1 1 1 2 2 , dumbCut(1:7,3) is 1 1 1 2 2 2 3 and dumbCut(1:7,thre=c(0,2,3,5)) is 1 1 2 3 3 4 4 .

Next step is to convert the output to a factor (because using numbers for categories in R is simply a masochism) and generate a meaningful level names, so basically replicating actual cut .