An error I can't understand. "Promise already under evaluation..."

Question

I am trying to write a function that finds pattern in names, with the help of stringr package. My function looks like following:

namezz=function(thepatternx,data=data,column=Name){

  library(stringr)

  thepattern=as.character(quote(thepatternx))

  pattern <- thepattern
  strings <- data$column ##data$column is a character vector
  found=str_detect(strings, pattern)
  yez= rownames(data[which(found==TRUE),])
  hhh=as.numeric(yez)+1
  return(hhh)

}

When I call the function I get an error:

namezz(Primus)

Error in namezz(Primus) : 
  promise already under evaluation: recursive default argument reference or earlier problems?

Can't understand the error, and what I am doing wrong.. Thanks in advance for any guidelines:)

EDIT: If I instead write it like this:

namezz=function(thepatternx,data,Name){

  library(stringr)

  thepattern=as.character(quote(thepatternx))

  pattern <- thepattern
  strings <- data$Name  #####data$column is a character vector
  found=str_detect(strings, pattern)
  yez= rownames(data[which(found==TRUE),])
  hhh=as.numeric(yez)+1
  return(hhh)

}

I get:

namezz(Primus,data,Name)

numeric(0)

Which is not correct, because if I do the procedure without putting it in a function I get the rows:

pattern="Primus"
strings <- data$Name
mja=str_detect(strings, pattern)
yez= rownames(data[which(mja==TRUE),])
hhh=as.numeric(yez)+1

    [1] 2 3 4 5 6 7 8 9

Here is a dput:

dput(head(data))
structure(list(Year = 1901:1906, Name = c(">>Primus<< sbk", ">>Primus<< sbk", 
">>Primus<< sbk", ">>Primus<< sbk", ">>Primus<< sbk", ">>Primus<< sbk"
), Established = c(1899L, 1899L, 1899L, 1899L, 1899L, 1899L), 
    Bolagskod = c(2L, 2L, 2L, 2L, 2L, 2L), Kategori = c(2L, 0L, 
    0L, 0L, 0L, 0L), BranschTillhörighet = c(2L, 2L, 2L, 2L, 
    2L, 2L), Startår = c(1901L, 1901L, 1901L, 1901L, 1901L, 1901L
    ), Stoppår = c(1908L, 1908L, 1908L, 1908L, 1908L, 1908L), 
    Ranges = c("8  1901 - 1908  >>Primus<< sbk", "8  1901 - 1908  >>Primus<< sbk", 
    "8  1901 - 1908  >>Primus<< sbk", "8  1901 - 1908  >>Primus<< sbk", 
    "8  1901 - 1908  >>Primus<< sbk", "8  1901 - 1908  >>Primus<< sbk"
    ), Years.present = c("1901  1902  1903  1904  1905  1906  1907  1908", 
    "1901  1902  1903  1904  1905  1906  1907  1908", "1901  1902  1903  1904  1905  1906  1907  1908", 
    "1901  1902  1903  1904  1905  1906  1907  1908", "1901  1902  1903  1904  1905  1906  1907  1908", 
    "1901  1902  1903  1904  1905  1906  1907  1908"), Delägare.män. = c(267L, 
    271L, 317L, 339L, 339L, 345L), Delägare.kvinnor. = c(246L, 
    251L, 236L, 244L, 260L, NA), Sjukdomsfall.män. = c(66L, 61L, 
    100L, 103L, 106L, 82L), Sjukdomsfall.kvinnor. = c(59L, 55L, 
    60L, 71L, 85L, 60L), Sjukdagar.män. = c(1686L, 1918L, 2149L, 
    2212L, 2331L, 1890L), Sjukdagar.kvinnor. = c(1681L, 1197L, 
    1589L, 1904L, 2282L, 1750L), Inkomster.InträdesAvgifter. = c(303L, 
    NA, NA, NA, NA, NA), Inkomster.RegelbundnaAvgifter. = c(4901L, 
    4939L, 5172L, 5687L, 5728L, 5879L), Inkomster.UtdebiteradeAvgifter. = c(1313L, 
    1045L, 1141L, 2024L, 1462L, 1934L), Inkomster.Böter. = c(241L, 
    NA, NA, NA, NA, NA), SummaMedl.avg. = c(6758L, 5984L, 6313L, 
    7711L, 7190L, 7813L), Inkomster.BidragStatKommun. = c(366L, 
    440L, 456L, 464L, 476L, 493L), Inkomster.Räntor. = c(NA_integer_, 
    NA_integer_, NA_integer_, NA_integer_, NA_integer_, NA_integer_
    ), Inkomster.Övrigt. = c(24L, 722L, 874L, 605L, 805L, 647L
    ), Inkomster.Summa. = c(7148L, 7146L, 7644L, 8781L, 8472L, 
    8954L), DiffIntäkter.SummaMotVerkligSumma. = c(0L, 0L, -1L, 
    -1L, -1L, -1L), Utgifter.Sjukhjälp. = c(4735L, 4450L, 5300L, 
    5870L, 6560L, 5200L), Utgifter.Begravningshjälp. = c(1200L, 
    795L, 1045L, 1810L, 955L, 1675L), Utgifter.Arvoden. = c(NA_integer_, 
    NA_integer_, NA_integer_, NA_integer_, NA_integer_, NA_integer_
    ), Utgifter.Förvaltning. = c(956L, 972L, 1038L, 1156L, 1523L, 
    1171L), Utgifter.Övrigt. = c(25L, NA, 20L, 5L, NA, NA), Utgifter.Behållning. = c(231, 
    929, 240, -59, -565, 908), Utgifter.SummaÖvrigt.Behållning. = c(256L, 
    929L, 260L, -54L, -565L, 908L), Utgifter.Summa. = c(7148L, 
    6217L, 7403L, 8841L, 9038L, 8046L), KOLL = c(-1L, 0L, 0L, 
    0L, 0L, 0L), Tillgångar.KontantIKassa. = c(835L, 1765L, 2006L, 
    1946L, 1380L, 2259L), Tillgångar.KontantMedelBank. = c(NA_integer_, 
    NA_integer_, NA_integer_, NA_integer_, NA_integer_, NA_integer_
    ), Tillgångar.Totalt. = c(836L, 1765L, 2006L, 1946L, 1468L, 
    2348L), Skulder.Totalt. = c(NA_integer_, NA_integer_, NA_integer_, 
    NA_integer_, NA_integer_, NA_integer_), TillgångarÖverSkulder = c(NA_integer_, 
    NA_integer_, NA_integer_, NA_integer_, NA_integer_, NA_integer_
    )), .Names = c("Year", "Name", "Established", "Bolagskod", 
"Kategori", "BranschTillhörighet", "Startår", "Stoppår", "Ranges", 
"Years.present", "Delägare.män.", "Delägare.kvinnor.", "Sjukdomsfall.män.", 
"Sjukdomsfall.kvinnor.", "Sjukdagar.män.", "Sjukdagar.kvinnor.", 
"Inkomster.InträdesAvgifter.", "Inkomster.RegelbundnaAvgifter.", 
"Inkomster.UtdebiteradeAvgifter.", "Inkomster.Böter.", "SummaMedl.avg.", 
"Inkomster.BidragStatKommun.", "Inkomster.Räntor.", "Inkomster.Övrigt.", 
"Inkomster.Summa.", "DiffIntäkter.SummaMotVerkligSumma.", "Utgifter.Sjukhjälp.", 
"Utgifter.Begravningshjälp.", "Utgifter.Arvoden.", "Utgifter.Förvaltning.", 
"Utgifter.Övrigt.", "Utgifter.Behållning.", "Utgifter.SummaÖvrigt.Behållning.", 
"Utgifter.Summa.", "KOLL", "Tillgångar.KontantIKassa.", "Tillgångar.KontantMedelBank.", 
"Tillgångar.Totalt.", "Skulder.Totalt.", "TillgångarÖverSkulder"
), row.names = c(NA, 6L), class = "data.frame")

Edit

This works:

namezz=function(thepatternx,data,Name){

  library(stringr)

  thepattern=thepatternx

  pattern <- thepattern
  strings <- data$Name
  mja=str_detect(strings, pattern)
  yez= rownames(data[which(mja==TRUE),])
  hhh=as.numeric(yez)+1
  return(hhh)  

} 

namezz("Primus",data,Name)
[1] 2 3 4 5 6 7 8 9

But how can i pass Primus without the quotation marks, namezz(Primus,data,Name)?. Thinking about something as in my question, but as.character(quote()) does not work..

Whoa. Be a little patient and give people a chance to have a look and see if they can know what is going on! It takes time for people to understand the problem and formulate a response good enough to earn those well deserved rep points!! — Simon O'Hanlon, Feb 20 '13 at 14:58
The problem is your use of `data=data`, as explained in the linked question. Fix that first, then debug the remaining bugs in your code. — Andrie, Feb 20 '13 at 15:00
FYI `data[which(found==TRUE),]` is equivalent to `data[found, ]`, and you should use `data[found, , drop = FALSE]` to avoid problems if you pass in a one-col data frame. — hadley, Feb 20 '13 at 22:35
Although maybe you actually want `which(found) + 1` instead of `as.numeric(yez)+1` — hadley, Feb 20 '13 at 22:37

Simon O'Hanlon · Accepted Answer · 2013-02-20T15:20:08.753

Someone can correct me on this, but I think that you are passing Primus and Name as objects to the function and it is looking in the .GlobalEnv for those objects and is not finding them, therefore your function is failing to carry out most of your instructions (and is returning nothing). I have edited your function a bit.

Instead try this...

 namezz <- function( pattern = " ", data , column= "Name" ){
   library(stringr)
   strings <- data[ , column ] ##data$column is a character vector
   found = str_detect( strings , pattern )
   yez = rownames( data[ which( found==TRUE ) , ] )
   hhh = as.numeric( yez ) + 1
   return( hhh )
 }

Then you must use the function like so:

namezz( "Primus" , data = data ) #In this case the default for column is "Name" as you want

The problem with passing data = data is explained very nicely here. An excerpt from that post (where they refer to testparams you would refer to data)...

"One of the most important things to know about the evaluation of arguments to a function is that supplied arguments and default arguments are treated differently. The supplied arguments to a function are evaluated in the evaluation frame of the calling function. The default arguments to a function are evaluated in the evaluation frame of the function."

the parameter testparams, when no matching argument is passed, is given the default value which is the value of the variable testparams looked-up not in the environment where foo is defined, and not in the environment where foo is called, but rather in the local environment created when the function is called and where parameters are mapped to values -- and in this environment, testparams is a parameter, which is already being under evaluation, hence the recursive lookup error.

thanks I managed that too. Perhaps I was a bit unclear. I want to pass `Primus` without the quotation marks...Any idea? See my edit in the end of the question. Have a good day! — user1665355, Feb 20 '13 at 15:20
Please explain why you have to pass it without quotation marks? AFAIK this is not possible. R will search for the object called Primus in the calling frame of the function unless you surround it with quotation marks. There are people with more experience and knowledge than me that may correct me. — Simon O'Hanlon, Feb 20 '13 at 15:22
Ok thanks! Good answer:) But I think I saw somewhere a post where there was a way around it.. But I might be wrong!!:) — user1665355, Feb 20 '13 at 15:32
Thanks. I am now interested to know if this is possible! @hadley might have a clue... — Simon O'Hanlon, Feb 20 '13 at 15:39

An error I can't understand. "Promise already under evaluation..."

1 Answers1

Linked