I have a table with timestamp and 9 columns filled with Numbers. Default, they were characters so I used:
library(dplyr)
data[-1] <- data[-1] %>% mutate_all(~as.numeric(as.character(.)))
Which worked just fine without any error. Now i want to group them and turn each value into "-", "o", "+" but it doesn't work, somehow...
for(j in 2:10) data[,j] <- cut(data[,j], c(-4,-0.5,0.5,4), labels= c("-","o","+"))
The error that appears (I translated myself, original in german) "Fehler in cut.default(aggr_beer1[, j], c(-4, -0.5, 0.5, 4), labels = c("-", : 'x' muss numerisch sein" "Error in cut.default(aggr_beer1[, j], c(-4, -0.5, 0.5, 4), labels = c("-", : 'x' must be numeric"
Can someone please help me? What am I doint wrong?
Notice the difference between
library(dplyr)
data[,2]
# A tibble: 5 × 1
a
<dbl>
1 -3
2 -2
3 -1
4 0
5 1
and
as.data.frame(data)[,2]
[1] -3 -2 -1 0 1
Subsetting a tibble by column (data[,X]) returns a tibble but cut needs a numeric vector. A data.frame drops the data frame structure when subsetting by only one column. As a side note, to keep the data frame use drop=F, e.g. as.data.frame(data)[,2,drop=F]
Another way to approach this, if this is your data
data <- structure(list(grp = c("A", "B", "C", "D", "E"), a = c("-3",
"-2", "-1", "0", "1"), b = c("2", "3", "4", "5", "6")), row.names = c(NA,
-5L), class = c("tbl_df", "tbl", "data.frame"))
data
# A tibble: 5 × 3
grp a b
<chr> <chr> <chr>
1 A -3 2
2 B -2 3
3 C -1 4
4 D 0 5
5 E 1 6
Get the desired result with across
library(dplyr)
data %>%
mutate(across(a:b, ~
cut(as.numeric(.x), c(-4,-0.5,0.5,4), labels= c("-","o","+"))))
# A tibble: 5 × 3
grp a b
<chr> <fct> <fct>
1 A - +
2 B - +
3 C - +
4 D o NA
5 E + NA
Note that mutate_all has been superseded by across, see ?mutate_all.
Use for(j in 2:3) data[,j] <- cut(data[,j,drop=T], c(-4,-0.5,0.5,4), labels= c("-","o","+")) to get your approach working.