Accessing Dataframe columns using bracket vs dot notation in Julia

Question

Which is the recommended way to access Dataframe columns in Julia: Using df[:col] or df.col? For example, in Pandas Dataframe, the recommended way is using the bracket notation for performance as well as other reasons as discussed here: Speed difference between bracket notation and dot notation for accessing columns in pandas.

So, is there a similar recommendation for using bracket notation over the dot notation for Dataframes in Julia, either for performance or otherwise? Or are both equally good?

Answer 1

They are equally good, and normally df.col is used. There are two cases when you need to use the other style:

• when column name is not a valid identifier, e.g. df[Symbol("Some column name with spaces")]
• soon broadcasting for data frames will be implemented and then if col does not exist in df the syntax df.col .= 1 will not work while df[:col] .= 1 will create a new column.

Note that df.col notation is handy as if you write df. and press <tab> in REPL or Jupyter Notebook you will get a list of column names.