ERROR: This tidyselect interface doesn't support predicates yet.

[MartinEr]

To whom it may concern,

Hello, this is Martin.

I am new in R and I am having a problem with the function unite in the package tidyverse.

I am trying to unite three variables (day, month, year) into a newone (birth_date) that would display dd_mm_yyyy. I use the function unite but I have a problem with missing values. It gives me:

Proof1 <- unite(Proof1, birth_date, day:year, sep = "_", remove = FALSE, na.rm = TRUE)
Errore: This tidyselect interface doesn't support predicates yet.
Contact the package author and suggest using eval_select().
Run rlang::last_error() to see where the error occurred.

If I do set na.rm=FALSE, then when there are missing values in the three initial columns I want to unite, it just creates something like this: "NA_NA_NA", so it does not identify missing values. I would like those values to appear as missing values.

Anyone has a solution for this?
Thanks very much!

Mar

[martin.R]

Try this change:

Proof1 <- unite(Proof1, "birth_date", day:year, sep = "_", remove = FALSE, na.rm = TRUE)

[MartinEr]

Thank you martin.R!

I tried but it still gives me the same error. I guess it is related with missing values. Any other idea?

Thanks again,

Mar

[martin.R]

You'll need to provide your example data in that case via e.g. dput().

[nirgrahamuk]

Does the following code give you a tidyselect error ?

library(tidyverse)
df <- data.frame(day=c(1,2,3),
                 month=c(7,NA,9),
                 year=c(2020,2020,2020),
                 somedata=LETTERS[1:3])

df %>% unite(created_date,day:year,remove = FALSE)

[MartinEr]

Hi,

thank you for your answers. I try to reformulate the problem (I am not very familiar with dput(), sorry, I am very new with R). Here is an example dataset.

I have columns Day, Month and Year and I want to create variable birth_date. When I have missing values in at least one of the three columns (Day, Month, Year), I'd like to have a missing value in the new variable as well.

When I use the function:
Proof1 <- unite(Proof1, birth_date, day:year, sep = "_", remove = FALSE, na.rm = TRUE)
I get the following error:
Errore: This tidyselect interface doesn't support predicates yet.
Contact the package author and suggest using eval_select() .
Run rlang::last_error() to see where the error occurred.

I hope this is a bit clearer. Thank you for your help!

Mar

[MartinEr]

Hi,

Thanks for helping.
The code does not give me the error any more, but it does not solve the problem with missing values, plus it creates the variable in a new dataset.
Any other idea?

Mar

[nirgrahamuk]

run your code up to the line before you would use unite().
then run

dput(head(Proof1,n=20))

and paste the results here.

[MartinEr]

@nirgrahamuk

dput(head(Proof1,n=20))

here it is:
structure(list(day = c(NA, "01", "02", "08", NA, "01", "02",
"08", NA, "01", "02", "08", NA, "01", "02", "08", NA, "01", "02",
"08"), month = c(NA, "03", "09", "09", NA, "03", "09", "09",
NA, "03", "09", "09", NA, "03", "09", "09", NA, "03", "09", "09"
), year = c(NA, "2008", "2008", "2010", NA, "2008", "2008", "2010",
NA, "2008", "2008", "2010", NA, "2008", "2008", "2010", NA, "2008",
"2008", "2010")), row.names = c(NA, -20L), class = c("tbl_df",
"tbl", "data.frame"))

thanks,

Mar

[nirgrahamuk]

if you use the lubridate package, it can transform text representations of dates to R date dtypes, failures will get an NA.

library(tidyverse)
library(lubridate)
df2 <- df %>% 
  unite(created_date_txt,day:year,remove = FALSE) %>% 
  mutate(created_date = lubridate::dmy(created_date_txt))

[MartinEr]

Thanks nirghrahamuk.
When implementing your suggestion it gives me this error:
Error in UseMethod("unite_") :
no applicable method for 'unite_' applied to an object of class "function"

[nirgrahamuk]

df should be the name of your input dataframe

[MartinEr]

Thanks for your help. It was really useful.