Be advised: this thread is not for serious in-depth discussion of weighty topics (we have a link for that), this thread is not for anything Culture War related. This thread is for Fun. You got jokes? Share 'em. You got silly questions? Ask 'em.
- 145
- 1
Jump in the discussion.
No email address required.
Notes -
Far and away the most frustrating ones are when I misspell a variable exactly once in an awkward spot.
Although, R has this thing where the logicals (booleans) TRUE and FALSE can also be replaced with T and F for conciseness (you can pass something like ‘param=T’ without typing the whole thing, which is awesome) … but T and F unlike TRUE and FALSE are not reserved key words! Once I accidentally assigned (capital) T or F as a variable and that was a nightmare to figure out why random things were breaking all over.
The other infrequent but annoying thing is “factors”. In data analysis often you have categorical variables with preset and limited values. You can make these into “factors” but really it’s stored as a named integer vector. Say we are storing HTML codes, 200, 404, whatever. You might in some common analyses cases treat them as strings, and categorical, although obviously they are maybe still ‘numbers’ in your head. If you aren’t careful, you might accidentally write ‘as.numeric(HTML_codes)’ but that will return the integers that are used as internal representations, (like 1,1,2,4,1) rather than the numbers (really strings) themselves. Because they aren’t strings anymore, they are factors as a data type, and R decided rather than have a special unique data type they would just implement it as an integer with metadata (the actual “value” eg “200” is stored as a “name” attribute). Honestly there are good reasons for that, but still a major gotcha. (A very common package import makes this a lot less mistake prone but sometimes coding in a hurry you might forget)
More options
Context Copy link