Problem when function argument is the same as dataframe column name in dplyr

[Weicheng Zhu]

Hi all,

I have a problem that bothered me for a long time and I would like to

hear advices on how to handle this elegantly.

The problem occurs when the argument of a function passed to the `filter` function

happens to be the same as a column name in the dataframe as shown in the following

example. I initially want to filter based on the user input value of the `doy`, but apparently

`dplyr` uses both `doy` in the `filter` function from the data frame.

My current way is to rename the function parameter as shown in `f2` to avoid using the

same names that occur in the dataframe. But I really don't like the syntax `doy1`.

I'm wondering whether there is a better way to handle this. 

Thanks for any suggestions!

df = data.frame(year = rep(rep(c(2010:2011), each = 3)), 

                doy = c(3,200, 365, 3, 100, 365))

f1 = function(df, yr, doy){

  df %>% filter(year == yr, doy == doy) ## Problem

}

f1(df, 2010, 3)

#   year doy

# 1 2010   3

# 2 2010 200

# 3 2010 365

f2 = function(df, yr, doy1){

  df %>% filter(year == yr, doy == doy1) 

}

f2(df, 2010, 3)

# year doy

# 2010   3

Best,

Weicheng

[Maximilian]

Hi,

you simply have to use local() for the variable you pass into the function:

f1 = function(df, yr, doy){

  browser()

  df %>% filter(year == yr, doy == local(doy))

}

Best

Max

[Weicheng Zhu]

Thanks Max! This is the solution I want.