DataFrames Cheatsheet review about www.ahsmart.com HOT 3 CLOSED

The reason for using sibsp and parch for stack is that they have the same data type. I don't really care much about the context since the main purpose is to remind people about the syntax. Just curious, how much are you bothered by this?

under "Select observations" section, you might want to mention that you can use unique! to mutate to be consistent.

👍

same section - I often use unique!(df, [cols...]) syntax, which only checks certain columns for uniqueness. Might be worth adding

Interesting. I don't know about this syntax before. It seems to be a somewhat "dangerous" usage since some data get lost? Under what circumstances do you want to use this feature?

Not totally clear what that weird "R" symbol - indexing means. Is it just row-wise? Or related to R language?

Haha... good catch. I was in a hurry to make it look nice and didn't pay attention to what it means. I think it came to my mind because in this simple syntax df[1,2] for which the indices are just numbers. Any suggestion?

Also, that syntax can also be used to index columns at the same time - is that worth a note? (I realize space is limited)

Yes, I can add.

from www.ahsmart.com.

See updated cheat sheet at #3

from www.ahsmart.com.

Just curious, how much are you bothered by this?

Not a lot, just a minor stumbling block. From a usability perspective, it would just be more intuitive if I didn't have to wonder about the data type.

Interesting. I don't know about this syntax before. It seems to be a somewhat "dangerous" usage since some data get lost? Under what circumstances do you want to use this feature?

Yeah, can lose data, but sometimes I want this. For example, I have a longitudinal dataset of subjects, and I just want to use the first time-point from each subject to do some cross-sectional analysis, so I do

sort!(df, :timepoint)
crosssection = unique(df, :subject)

Any suggestion?

Nope, sorry :-)

from www.ahsmart.com.