Why does var act like cov in R? -

Sorry to ask ... this is definitely a common question, and this is a silly question, but It is annoying me to assume that I want to get the variance of each numerical column in a data frame, like

  df < - data.frame (x = 1: 5, y = seq (1,50,10)))   
 Naturally, I'll try  
 < Code> var (df)   
 Instead of giving me what I hope, what will happen like this  
  xy 2.5 250   
 I get it  
  xyx 2.5 25 y 25.0 250   
 which are variations in diagonals, and other places have correlation . Which makes sense when I look at the search (var) and read "var is another interface for the COV", the difference is a convertible and the flow of between yourself, of course. Output is a bit misleading, but I can read with diagonals, or can only generate variants using  diag (var (df)) ,  sapply (df, var) , or repeatedly on  lapply (df, var) , or  var  and  df $ x  and  df $ y  Code>.  
 But why? Variance is a regular, basic descriptive figure, which means only second. Should not it be completely and completely trivial to implement in the data frame column? When I was asked only for difference, why should I sympathize? just curious. Thanks for any comments on this.   
 
 
 idiomatic approach  
  Sapply (df, var)   
  var   has a method for  data.frames  with  data.frame  By supporting a  matrix .  
  variation  is a regular basic descriptive figure, so correlation and correlation are they are all interconnected and interesting, especially if you use a linear model Want to  
 You want you to always be able to accomplish your own work  
  Var & lt; - Function (x, ...) {if (is.data.frame (x)) {Return (sapply (x, var, ...) else {Return (var (x, ...))}}