There are some values in my dataset(df) that needs to be replaced with correct values e.g.,
| Height | Disease | Weight>90kg |
|---|---|---|
| 1.58 | 1 | 0 |
| 1.64 | 0 | 1 |
| 1.67 | 1 | 0 |
| 52 | 0 | 1 |
| 67 | 0 | 0 |
I want to replace the first three values as '158', '164' & '167'. I want to replace the next as 152 and 167 (adding 1 at the beginning).
I tried the following code but it doesn't work:
data_clean <- function(df) {
df[height==1.58] <- 158
df}
data_clean(df)
Please help!
Using recode you can explicitly recode the values:
df <- mutate(df, height = recode(height,
1.58 = 158,
1.64 = 164,
1.67 = 167,
52 = 152,
67 = 167))
However, this obviously is a manual process and not ideal for a case with many values that need recoding.
Alternatively, you could do something like:
df <- mutate(df, height = case_when(
height < 2.5 ~ height * 100,
height < 100 ~ height + 100
)
This would really depend on the makeup of your data but for the example given it would work. Just be careful with what your assumptions are. Could also have used is.double and 'is.integer`.
mutate(df, height = case_when(height == 1.58 ~ 158, height == 1.64 ~ 164, ...case_when you need to explicitly name all conditions otherwise the other values will be NA. The typical way to do this is add TRUE ~ height do the end of your case_when. In other words, if you reach this condition, simply return the actual value.You also could vectorize a switch:
foo <- Vectorize(FUN = function(x) {
switch(as.character(x),
"1.58" = 158,
"1.64" = 164,
"1.67" = 167,
"52" = 152,
"67" = 167,
NA)})
Then just replace as follows:
df$Height <- foo(df$Height)