I have a df and I want to do a simple calculation using test*factor. How could I take care of the signs in front of the value?
df<-structure(list(test = structure(c(3L, 2L, 5L, 4L, 1L), .Label = c("<=1.5",
"<5", ">=1", ">7.8", "6"), class = "factor"), factor = c(2, 4,
6, 9, 8)), row.names = c(NA, -5L), class = "data.frame")
Another possible solution:
library(tidyverse)
df %>%
mutate(outcome = str_c(str_remove(test, "\\d+"), factor *
as.numeric(str_extract(test, "\\d+"))))
#> test factor outcome
#> 1 >=1 2 >=2
#> 2 <5 4 <20
#> 3 6 6 36
#> 4 >78 9 >702
#> 5 <=15 8 <=120
Or with parse_number:
library(tidyverse)
df %>%
mutate(outcome = str_c(str_remove(test, "\\d+"),
factor * parse_number(test %>% as.character())))
#> test factor outcome
#> 1 >=1 2 >=2
#> 2 <5 4 <20
#> 3 6 6 36
#> 4 >78 9 >702
#> 5 <=15 8 <=120
With decimals:
library(tidyverse)
df %>%
mutate(outcome = str_c(str_remove(df$test, "\\d+(\\.\\d+)?"),
factor * parse_number(test %>% as.character())))
#> test factor outcome
#> 1 >=1 2 >=2
#> 2 <5 4 <20
#> 3 6 6 36
#> 4 >7.8 9 >70.2
#> 5 <=1.5 8 <=12
We can try
transform(
df,
outcome = paste0(gsub("\\d","",test),as.numeric(gsub("\\D", "", test)) * factor)
)
which gives
test factor outcome
1 >=1 2 >=2
2 <5 4 <20
3 6 6 36
4 >78 9 >702
5 <=15 8 <=120
One more. Here we use parse_number to multiply and then paste after removing the numbers from test column:
library(dplyr)
library(readr)
df %>%
mutate(outcome = paste0(str_replace(test, '\\d+', ''), factor*parse_number(as.character(test))
))
test factor outcome
1 >=1 2 >=2
2 <5 4 <20
3 6 6 36
4 >78 9 >702
5 <=15 8 <=120
We could use str_replace - in the pattern, match one or more digits (\\d+) and replace with a function to multiply with the factor column after converting to numeric
library(stringr)
df$outcome <- str_replace(df$test, "\\d+", function(x) as.numeric(x) * df$factor)
-output
df$outcome
[1] ">=2" "<20" "36" ">702" "<=120"
For the updated, change the \\d+ to include the decimal as well
str_replace(df$test, "[0-9.]+", function(x) as.numeric(x) * df$factor)
[1] ">=2" "<20" "36" ">70.2" "<=12"
And another solution using separate:
library(tidyr)
library(stringr)
library(dplyr)
df %>%
separate(test,
into = c("temp1", "temp2"),
sep = "(?<=\\D|^)(?=\\d+)",
remove = FALSE) %>%
mutate(
temp3 = as.numeric(temp2)*factor,
outcome = str_c(temp1, temp3)
) %>%
select(-matches("temp"))
test factor outcome
1 >=1 2 >=2
2 <5 4 <20
3 6 6 36
4 >78 9 >702
5 <=15 8 <=120
Here is an option using data.table and stringr:
library(data.table)
dt <- as.data.table(df)
dt[, outcome := str_c(str_remove(test, "[0-9.]+"), as.numeric(str_extract(test, "[0-9.]+")) * factor)]
Output
test factor outcome
1: >=1 2 >=2
2: <5 4 <20
3: 6 6 36
4: >7.8 9 >70.2
5: <=1.5 8 <=12
Or another option using tidyr::extract:
library(tidyverse)
df %>%
tidyr::extract("test", c("sign", "outcome"), "(\\D*)(\\d.*)", remove = F, convert = T) %>%
mutate(outcome = paste0(sign, outcome * factor)) %>%
select(names(df), outcome)
