Last year when I was at JSM, I attended an excellent talk by Nicholas Horton and later nitpicked his R code:
ifelse(dest == "ORD" | dest == "MDW", TRUE, FALSE)
There is no need to use
ifelse() here. The code above is equivalent to:
dest == "ORD" | dest == "MDW"
I have a feeling that some people are not comfortable when they think about Boolean values without literally seeing the values
FALSE. I think this is understandable. Personally I try to avoid redundancy, unless I’m explicitly thinking of the Department of Redundancy Department.
I also often see people write code like
if (x == TRUE). The
== TRUE part is redundant in my eyes:
if (x) is enough. I think there is a slight visual benefit of explicitly comparing
TRUE, though. That is, when glancing at the code, you can quickly know there is a Boolean comparison here, because the keyword
TRUE is visually prominent.
A few minutes ago I saw yet another example of redundancy in the article “Flawed code served as icing on the cake for Prime Minister’s social agency launch”. The last line of the R code on the cake was like:
minister_greeting = ifelse(welcomeSpeech == TRUE, "....")
This line was quite interesting in several aspects:
It is flawed because the third argument of the
ifelse()function was missing (with no default).
Again, there is redundancy in
welcomeSpeech == TRUE:
ifelse(welcomeSpeech)suffices. However, I think this redundancy is understandable since the cake is for someone who (I assume) do not really know much about R, and
ifelse(welcomeSpeech == TRUE)may make the code a little more readable to them.
The author of this piece of code used the left arrow
<-for assignments on all previous lines, but changed to the equal sign
=on the last line. As you all probably know, I use the euqal sign all the time (except when I collaborate with someone who insists on using the arrow), so I’d like to say “Welcome to the dark side and long live the equal sign!”