Replacing range of numeric values with a string only replaces some values-CodePudding

I have a data frame (BRAT2) where a column (oCC_HPE) with 58110 entries contains a range of values from 0-40 out to 2 decimal places.

When I try to reassign the range of values to a string all the values replace properly except for values from 5.02-9.99.

I'm not sure what is causing this issue, I've tried altering the order of the replacements, changing how many decimal places are in my replacement criteria, but to no avail.

BRAT2$oCC_HPE[BRAT2$oCC_HPE == 0.00] <- 'None'
BRAT2$oCC_HPE[BRAT2$oCC_HPE > 0.00 & BRAT2$oCC_HPE <= 1.00] <- 'Rare'
BRAT2$oCC_HPE[BRAT2$oCC_HPE > 1.00 & BRAT2$oCC_HPE <= 5.00] <- 'Occasional'
BRAT2$oCC_HPE[BRAT2$oCC_HPE > 5.00 & BRAT2$oCC_HPE <= 15.00] <- 'Frequent'
BRAT2$oCC_HPE[BRAT2$oCC_HPE > 15.00 & BRAT2$oCC_HPE <= 40.00] <- 'Pervasive'

CodePudding user response：

Instead of multiple assignments (after the first assignment on the same column, the column type is changed from numeric to character, making the next comparisons returning incorrect output), use either case_when or cut

BRAT2$oCC_HPE <- cut(BRAT2$oCC_HPE, breaks = c( 0, 1, 5, 15, 40, Inf),
    labels = c("None", "Rare", "Occasional", "Frequent", "Pervasive"))

Or use case_when from dplyr

library(dplyr)
BRAT2 <- BRAT2 %>%
     mutate(oCC_HPE = case_when(oCC_HPE == 0.00 ~ "None",
            oCC_HPE > 0.00 &oCC_HPE <= 1.00 ~ "Rare",
           oCC_HPE > 1.00 &oCC_HPE <= 5.00 ~  "Occasional",
          oCC_HPE > 5.00 &oCC_HPE <= 15.00 ~ "Frequent" ,
          oCC_HPE > 15.00 &oCC_HPE <= 40.00 ~"Pervasive" ))