R: DPLYR package: bind_rows failing when calls a custom function

Question

Using DPLYR and TIDYR, I'm trying to create a tidy version of a dataset where rows can be missing depending on the data of certain columns. I created a function that returns the rows missing (by creating them with default data) in a new tbl_df(data.frame) (I unit-tested it and it works okay with specific data).

However, when calling it from 'bind_rows', I get the following error: Error in data.frame(a, b, c,...: Object 'A' not found.

For example, my data looks like this:

A        B        C        D        E        ...
a1       b1       c1       d1       e1       ...
a2       b2       c2       d2       e2       ...
...

My code looks like this:

data_tidy <- data %>%

    <some other functions to clean up like 'mutuate', 'filter', etc.> %>%

    brind_rows(myCustomFunction(A, B, C, D, E... ))

Any ideas what I'm doing wrong? I'm still new to R, DPLYR/TIDYR...

Note: If I remove the last call to 'bind_rows', the table is cleanup as expected with the proper A, B, C, etc. columns. I also use a 'for' loop in this specific scenario which I know might not be optimal but for now, I will work with this version so I can get it to work and then try to optimize my code (or vectorize).

Thanks!

Why are you trying to pass a function to bind_rows()? What do you think that's going to do? It would help if you made your problem reproducible with sample input data and expected output for that input. I'm guessing you're using the wrong function here. — MrFlick
Each row within the data describe a contract year for a player. However, the original data I used to clean/tidy up was stopping at 2018 for every player but some contract goes further (up to 2022 for example). The custom method creates the rows missing (in my example 4 rows would be created - 2019, 2020, 2021, 2022 with the relevant data for other variables). The method I'm calling is supposed to be executed to return those missing rows and then use bind_rows to add them to the original tidy data. Hopefully that makes sense. — Martin

Jthorpe Jthorpe · Accepted Answer · 2015-02-26T21:33:19

In your call to foo %>% brind_rows(myCustomFunction(A, B, C, D, E... )), myCustomFunction(A, B, C, D, E... ) is being called as an ordinary R function, whereas I think you'r expecting that it be evaluted within the context of a dplyr function as in mutate(x = myCustomFunction(A, B, C, D, E... )) where the arguments A, B, C, D, E would be replaced by fields from the data.frame that is passed as the implicit first argument thanks to the %>% operator.

In short, you need to call myCustomFunction(A, B, C, D, E... ) in such a way that the arguments are scoped correctly, such as:

data_tidy <- data %>% 
    <some other functions to clean up like 'mutuate', 'filter', etc.>

brind_rows(do.call(myCustomFunction,data_tidy))

R: DPLYR package: bind_rows failing when calls a custom function

1 Answers