Why is it not advisable to use attach in R and what should I use instead

Question

Let's pretend we have a data frame x with the columns employment and income in it. The commands x$job for the data in the job column and x$income for the data in the income column are typically used to refer to the data in the frame.

When referring to the same data, however, using the command attach(x) eliminates the need for the data frame's name and the $ symbol. As a result, in the R code, x$job becomes job and x$income becomes income.

The issue is that many R experts advise against using the attach() verb while programming in R.

What is the most important reason behind this? What should be substituted in its place?

Answer 1

There is one more option that applies to both character and numeric data:

ux[which.max(tabulate(match(x, ux)))] mode - function(x) ux - unique(x) ux
That can produce a 10M-integer vector and determine its mode on my puny little PC in approximately 30 seconds.

The above method adopts the same strategy as which.max and delivers the first value of the set of modes if your data source is likely to have multiple modes. Use this variation (provided by @digEmAll in the comments) to return all modes:

Tabulate (match (x, ux)) ux[tab == max(tab)] Modes - function(x) - ux - unique(x)
}