Skip to content

R CMD check says no visible binding for global variable

11 messages · Ben Bolker, @vi@e@gross m@iii@g oii gm@ii@com, Sorkin, John +3 more

Original
Naresh Gurbuxani
# 
I have written a function which returns an SQL query result as a data.frame.  Each column of data.frame is a variable not explicitly defined.

For every column name, R CMD check says ?no visible binding for global variable <name>. Status: 1 NOTE

Is it possible to tell R CMD check that these variables are OK?

Thanks,
Naresh

Sent from my iPhone
Ben Bolker
# 
This might be better for r-package-devel at r-project.org (since you're 
asking a question about package-checking).

   This is a common problem when using non-standard evaluation. 
Typically you can either use `utils::globalVariables` or set these 
variables to NULL near the top of your function.  There are arguments 
for either approach; I'm sure this has been discussed on r-package-devel 
before but I couldn't quickly locate a thread.
On 2025-01-27 5:46 p.m., Naresh Gurbuxani wrote:
@vi@e@gross m@iii@g oii gm@ii@com
# 
Naresh,

I am not sure how you are creating your data.frame so it has no, I think, column names. There are two scenarios including one where it is not really a valid data.frame and one where it can be handled before any other use as shown below. If it cannot be used, you might need to modify how your SQL or the function you call creates it so it includes either names it chooses or that you supply.

One silly solution if to give your data frame names before using it later. In prticulr, if you know what the columns contain, you can choose suitable names like this if you have exactly three columns:
first second third
1     1      2     3

If you have a varying number of columns and don't care what the names are, you can make n names that look like temp1, temp2, ... tempn like this:
[1] "temp1" "temp2" "temp3"

Obviously, you substitute in whatever your data.frame is called.

So the code to add names for columns looks like:
temp1 temp2 temp3
1     1     2     3


-----Original Message-----
From: R-help <r-help-bounces at r-project.org> On Behalf Of Naresh Gurbuxani
Sent: Monday, January 27, 2025 5:46 PM
To: r-help at r-project.org
Subject: [R] R CMD check says no visible binding for global variable

I have written a function which returns an SQL query result as a data.frame.  Each column of data.frame is a variable not explicitly defined.

For every column name, R CMD check says ?no visible binding for global variable <name>. Status: 1 NOTE

Is it possible to tell R CMD check that these variables are OK?

Thanks,
Naresh

Sent from my iPhone
______________________________________________
R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide https://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Sorkin, John
# 
There you go, once again helping strengthen ;)
John

Get Outlook for iOS<https://aka.ms/o0ukef>
Naresh Gurbuxani
# 
Data.frame is returned by SQL query.  It does have column names.  In the function, I make small changes to some columns.

Something like:

Myquery <- ?SELECT date, price, stock FROM stocktab WHERE stock = ?ABC? AND date > ?2025-01-01?;?

Prices <- dbGetQuery(con, myquery)
SetDT(Prices)
Prices[, date = as.Date(date)]


R CMD check say ?no visible binding for global variable ?date??


Sent from my iPhone
On Jan 28, 2025, at 1:24?AM, Sorkin, John <jsorkin at som.umaryland.edu> wrote:
?
There you go, once again helping strengthen ;)
John

Get Outlook for iOS<https://aka.ms/o0ukef>
Duncan Murdoch
#
On 2025-01-28 1:55 p.m., Naresh Gurbuxani wrote:
If Prices were a regular dataframe at this point, then the message would 
be correct. You can't calculate `as.Date(date)` without telling R where 
to look for the `date` variable.

However, you have set it to be a data.table instead.  They use 
nonstandard evaluation and look up `date` in the columns of `Prices`, 
and things work.  However, R's checks don't know this, so you still get 
the complaint.

The fix given by others is easiest:  sometime before this add a line

   date <- NULL

and it will satisfy the check code.

Duncan Murdoch
Naresh Gurbuxani
# 
This solution worked.
Thanks

Sent from my iPhone
@vi@e@gross m@iii@g oii gm@ii@com
# 
Naresh,

 

I think your details suggest your earlier request was not well defined.

 

Your problem looks like this line of code was wrong and did something you did not expect. It should be easy to fix. But I note you changed Prices to use the data.table package. Mind you, I thought the function was lower case setDT and you mention SetDT. At this point, I cannot help you as I do not know enough about the package.

 

As a general rule, the best way to ask for help includes mentioning the packages you used. The code snippet did not have a library() statement that would be needed.

 

Here is the original:

 

Prices[, date = as.Date(date)]

 

I cannot speak for how data.table might parse that. In base R, it would not work and I can suggest what would.

 

What you wanted to do was convert a column containing a date in some format into a date in some other format. Am I right? In particular, I am guessing the date returned by SQL into the Prices data.frame is there as text in a standard format. If not, you may need to convert the date some more specific way.

 

Here is a way outside data.table that might work within it as well.

 

Prices$date <- as.Date(Prices$date)

 

The above replaces the contents of one vector/column and note the suggested use is not of ?=? which gave you a side-effect, but to use ?<-? in the R-thritic way.

 

There are other variations and yet more if using the dplyr or other packages, but the above is fairly simple and straightforward and uses base R.

 

I will spare you further explanations and lecturing on what I thought your code would do as I looked again and realize it is not a standard data.frame anymore.

 

Here is a test case if your date is in this format:

 

Prices <- data.frame(date=c("2025-01-01", "2025-01-26"),

                     price=c(12, 14),

                     stock=c("AAPL", "MSFT"))

 

Prices$date <- as.Date(Prices$date)

 

The result is:
date price stock

1 2025-01-01    12  AAPL

2 2025-01-26    14  MSFT

 

 

If your date is in some other format, you have to specify the format. See the help on as.Date to see how to set the format=option.

 

If anyone wants to explain better on the data.table method and maybe explain what went wrong, I would be happy to hear it, as well as any corrections on what I wrote.

 

 

 

 

 

From: Naresh Gurbuxani <naresh_gurbuxani at hotmail.com> 
Sent: Tuesday, January 28, 2025 1:56 PM
To: John Sorkin <jsorkin at som.umaryland.edu>
Cc: avi.e.gross at gmail.com; r-help at r-project.org
Subject: Re: [R] R CMD check says no visible binding for global variable

 

Data.frame is returned by SQL query.  It does have column names.  In the function, I make small changes to some columns. 

 

Something like:

 

Myquery <- ?SELECT date, price, stock FROM stocktab WHERE stock = ?ABC? AND date > ?2025-01-01?;?

 

Prices <- dbGetQuery(con, myquery)

SetDT(Prices)

Prices[, date = as.Date(date)]

 

 

R CMD check say ?no visible binding for global variable ?date??

 

 

Sent from my iPhone
On Jan 28, 2025, at 1:24?AM, Sorkin, John <jsorkin at som.umaryland.edu <mailto:jsorkin at som.umaryland.edu> > wrote:
? 

There you go, once again helping strengthen ;)

John

 

Get Outlook for iOS <https://aka.ms/o0ukef> 

  _____  

From: R-help <r-help-bounces at r-project.org <mailto:r-help-bounces at r-project.org> > on behalf of avi.e.gross at gmail.com <mailto:avi.e.gross at gmail.com>  <avi.e.gross at gmail.com <mailto:avi.e.gross at gmail.com> >
Sent: Tuesday, January 28, 2025 12:01:25 AM
To: 'Naresh Gurbuxani' <naresh_gurbuxani at hotmail.com <mailto:naresh_gurbuxani at hotmail.com> >; r-help at r-project.org <mailto:r-help at r-project.org>  <r-help at r-project.org <mailto:r-help at r-project.org> >
Subject: Re: [R] R CMD check says no visible binding for global variable 

 

Naresh,

I am not sure how you are creating your data.frame so it has no, I think, column names. There are two scenarios including one where it is not really a valid data.frame and one where it can be handled before any other use as shown below. If it cannot be used, you might need to modify how your SQL or the function you call creates it so it includes either names it chooses or that you supply.

One silly solution if to give your data frame names before using it later. In prticulr, if you know what the columns contain, you can choose suitable names like this if you have exactly three columns:
first second third
1     1      2     3

If you have a varying number of columns and don't care what the names are, you can make n names that look like temp1, temp2, ... tempn like this:
[1] "temp1" "temp2" "temp3"

Obviously, you substitute in whatever your data.frame is called.

So the code to add names for columns looks like:
temp1 temp2 temp3
1     1     2     3


-----Original Message-----
From: R-help <r-help-bounces at r-project.org <mailto:r-help-bounces at r-project.org> > On Behalf Of Naresh Gurbuxani
Sent: Monday, January 27, 2025 5:46 PM
To: r-help at r-project.org <mailto:r-help at r-project.org> 
Subject: [R] R CMD check says no visible binding for global variable

I have written a function which returns an SQL query result as a data.frame.  Each column of data.frame is a variable not explicitly defined.

For every column name, R CMD check says ?no visible binding for global variable <name>. Status: 1 NOTE

Is it possible to tell R CMD check that these variables are OK?

Thanks,
Naresh

Sent from my iPhone
______________________________________________
R-help at r-project.org <mailto:R-help at r-project.org>  mailing list -- To UNSUBSCRIBE and more, see
https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fstat.ethz.ch%2Fmailman%2Flistinfo%2Fr-help <https://stat.ethz.ch/mailman/listinfo/r-help> &data=05%7C02%7CJSorkin%40som.umaryland.edu%7C25e7bd61380147526b8108dd3f58dced%7C717009a620de461a88940312a395cac9%7C0%7C0%7C638736373101270089%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=ExT4UObp9q1CqqqUNyr9GdX3fs6UFoCIndD0UFdQ2bE%3D&reserved=0
PLEASE do read the posting guide https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.r-project.org%2Fposting-guide.html <https://www.r-project.org/posting-guide.html> &data=05%7C02%7CJSorkin%40som.umaryland.edu%7C25e7bd61380147526b8108dd3f58dced%7C717009a620de461a88940312a395cac9%7C0%7C0%7C638736373101299265%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=KHzs8stxNBdUeRR9LW8DdYICbzLU3wovaKNJBqL%2BBOY%3D&reserved=0
and provide commented, minimal, self-contained, reproducible code.

______________________________________________
R-help at r-project.org <mailto:R-help at r-project.org>  mailing list -- To UNSUBSCRIBE and more, see
https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fstat.ethz.ch%2Fmailman%2Flistinfo%2Fr-help <https://stat.ethz.ch/mailman/listinfo/r-help> &data=05%7C02%7CJSorkin%40som.umaryland.edu%7C25e7bd61380147526b8108dd3f58dced%7C717009a620de461a88940312a395cac9%7C0%7C0%7C638736373101312257%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=ygouzLUqvgdAsZjnGVrFwzM93bYn%2Fbjevj5MyagGE04%3D&reserved=0
PLEASE do read the posting guide https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.r-project.org%2Fposting-guide.html <https://www.r-project.org/posting-guide.html> &data=05%7C02%7CJSorkin%40som.umaryland.edu%7C25e7bd61380147526b8108dd3f58dced%7C717009a620de461a88940312a395cac9%7C0%7C0%7C638736373101326801%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=kx%2FukXxxEEiVwui%2F6lIsA45q6Hiz97Qpj%2BwHi2eoSqA%3D&reserved=0
and provide commented, minimal, self-contained, reproducible code.
@vi@e@gross m@iii@g oii gm@ii@com
# 
That is an interesting fix Duncan suggested and it sounds now like everything WORKED as intended in data.table except that any checker being used externally is not able to del with things that look like a variable but are actually not a variable currently visible except within a function that is doing deferred evaluation. In this case, it recognizes the raw unevaluated string as being the name of a column within the data.table.

I would guess this can cause similar anomalies in the dplyr package when it does the same kinds of non-standard evaluation and I wonder even about the with() command or the within() which internally expand columns out so that commands using them work.

But setting the variable to NULL bothers me. Yes, it shuts up a check to see if the variable exists. But is there any reason you could not have a global variable and any number of local variables with the same name as the column, especially if non-standard evaluation only looks for the column and ignores other instances of the name in other environments? Setting it to NULL unconditionally could mess up some programs.

There are ways to set it conditionally and save the previous value and restore it after but this gets to be lots more work ...


-----Original Message-----
From: R-help <r-help-bounces at r-project.org> On Behalf Of Naresh Gurbuxani
Sent: Tuesday, January 28, 2025 4:25 PM
To: Duncan Murdoch <murdoch.duncan at gmail.com>
Cc: r-help at r-project.org
Subject: Re: [R] R CMD check says no visible binding for global variable

This solution worked.
Thanks

Sent from my iPhone
______________________________________________
R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide https://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Uwe Ligges
#
On 28.01.2025 22:52, avi.e.gross at gmail.com wrote:
The help pages of with() and within() do warn not to use it in package 
code ...
I'd avoid non standard evaluation in any case.

Best,
Uwe Ligges
Naresh Gurbuxani
# 
While my problem is solved, if not elegantly, data.table is widely used in packages.  

Any package builders using data.table who have encountered and solved this problem?

Sent from my iPhone