Writing function...receiving error that variable isn't found

Question

I can't figure out why I'm getting the error

Error: object 'grade' not found

when composing a function.

Grade is obviously in the dataset and is included in the function. If I don't use the function and just use

dat%>%
cohort.fun()%>%
group_by(cohort, variable, timepoint)%>%
    summarize(perf_measure = mean(measure))

everything works fine. These are the two functions I'm using:

library(reshape2)
library(tidyverse)

cohort.fun <- function(dat){
  dat%>%
    mutate(grade = as.numeric(grade))%>%
    mutate(cohort = ifelse(grade%in%c(3,4),3,ifelse(
      grade%in%c(5,6), 5, ifelse(
        grade%in%c(7,8), 7, grade))))%>%
    mutate(cohort = as.character(cohort))
}

melt.fun <- function(dat){
   melt(c("pid", "grade", "timepoint"), value.name = "measure")%>%
  cohort.fun()
}

then I run

dat%>%
melt.fun()

and I get the error above. Any ideas? Thanks much!

Here's the dput:

structure(list(pid = c("ADMIN-UCSF-bo004", "ADMIN-UCSF-bo005", 
"ADMIN-UCSF-bo008", "ADMIN-UCSF-bo010", "ADMIN-UCSF-bo011", "ADMIN-UCSF-bo012", 
"ADMIN-UCSF-bo013", "ADMIN-UCSF-bo014", "ADMIN-UCSF-bo015", "ADMIN-UCSF-bo016"
), grade = c("3", "3", "3", "3", "3", "3", "3", "3", "3", "3"
), RC1 = c(-1.81295211570392, -1.31252376878321, -1.1701654183369, 
-1.58244557144815, -1.95383829351231, -0.516109923323212, -0.370765686983851, 
-1.93212644807752, -1.6241046548069, -1.34160382084709), RC2 = c(-0.363819589341912, 
0.268206917949323, -2.24123725035034, -0.25274997192688, 0.313608190056975, 
-0.0393486670413662, -0.0623610937831014, 0.803692668734253, 
0.416065992573585, -0.069880541013785), RC3 = c(-2.69157047028032, 
-0.822917456389246, -1.52186068360016, -0.590070546800741, 0.583790188582597, 
-0.253888391947117, 1.22197349838073, -1.63335701437031, 1.24595192142446, 
0.0191275904777839), timepoint = c(1, 1, 1, 1, 1, 1, 1, 1, 1, 
1)), class = c("tbl_df", "tbl", "data.frame"), row.names = c(NA, 
-10L))

Yes, I've used that in my code and it works, but I'm trying to make a function within a function to simplify. Any idea why it doesn't work the way that I currently have it? — James, Apr 12 '20 at 20:25
I ask because then if I add to that: ```graph.fun <- function(dat){ group_by(cohort, variable, timepoint)%>% summarize(perf_measure = mean(measure)) }``` and tack that on to the code you have, I just get the same variation of error: ```Error in group_by(cohort, variable, timepoint) : object 'cohort' not found``` — James, Apr 12 '20 at 20:29
Thanks so much! Browser will be helpful. Stupid mistake....No, just trying to put cohort function within the melt.fun(). — James, Apr 12 '20 at 21:08

davidnortes · Accepted Answer · 2020-04-12T22:07:33.527

In your function melt.fun what you are actually passing to cohort.fun is the result of melt and not dat. Ergo cohort.fun does not find its parameter.

Defining melt.fun as follows:

melt.fun <- function(dat){
   melt(dat, c("pid", "grade", "timepoint"), value.name = "measure") %>%
  cohort.fun()
}

Should do the trick. You can see the use of the magrittr's forward pipe operators here

EDIT: I'm including the whole script here so you guys can see what I've done so far:

#Loading libraries
library(tidyverse)
library(reshape2)

#Loading data
dat <- structure(list(pid = c("ADMIN-UCSF-bo004", "ADMIN-UCSF-bo005", 
                              "ADMIN-UCSF-bo008", "ADMIN-UCSF-bo010", "ADMIN-UCSF-bo011", "ADMIN-UCSF-bo012", "ADMIN-UCSF-bo013", "ADMIN-UCSF-bo014", "ADMIN-UCSF-bo015", "ADMIN-UCSF-bo016"), grade = c("3", "3", "3", "3", "3", "3", "3", "3", "3", "3"), RC1 = c(-1.81295211570392, -1.31252376878321, -1.1701654183369, -1.58244557144815, -1.95383829351231, -0.516109923323212, -0.370765686983851, -1.93212644807752, -1.6241046548069, -1.34160382084709), RC2 = c(-0.363819589341912, 0.268206917949323, -2.24123725035034, -0.25274997192688, 0.313608190056975, -0.0393486670413662, -0.0623610937831014, 0.803692668734253, 0.416065992573585, -0.069880541013785), RC3 = c(-2.69157047028032, -0.822917456389246, -1.52186068360016, -0.590070546800741, 0.583790188582597, -0.253888391947117, 1.22197349838073, -1.63335701437031, 1.24595192142446, 0.0191275904777839), timepoint = c(1, 1, 1, 1, 1, 1, 1, 1, 1, 1)), class = c("tbl_df", "tbl", "data.frame"), row.names = c(NA, -10L))

#Defining functions
#Function1
cohort.fun <- function(dat){
  dat%>%
    mutate(grade = as.numeric(grade))%>%
    mutate(cohort = ifelse(grade%in%c(3,4),3,ifelse(
      grade%in%c(5,6), 5, ifelse(
        grade%in%c(7,8), 7, grade))))%>%
    mutate(cohort = as.character(cohort))
}

#Function2
melt.fun <- function(dat){
  melt(dat, c("pid", "grade", "timepoint"), value.name = "measure") %>%
  cohort.fun()
}

#Executing
dat%>%
  melt.fun()

#Result
> dat%>%
+   melt.fun()
                pid grade timepoint variable     measure cohort
1  ADMIN-UCSF-bo004     3         1      RC1 -1.81295212      3
2  ADMIN-UCSF-bo005     3         1      RC1 -1.31252377      3
3  ADMIN-UCSF-bo008     3         1      RC1 -1.17016542      3
4  ADMIN-UCSF-bo010     3         1      RC1 -1.58244557      3
5  ADMIN-UCSF-bo011     3         1      RC1 -1.95383829      3
6  ADMIN-UCSF-bo012     3         1      RC1 -0.51610992      3
7  ADMIN-UCSF-bo013     3         1      RC1 -0.37076569      3
8  ADMIN-UCSF-bo014     3         1      RC1 -1.93212645      3
9  ADMIN-UCSF-bo015     3         1      RC1 -1.62410465      3
10 ADMIN-UCSF-bo016     3         1      RC1 -1.34160382      3
11 ADMIN-UCSF-bo004     3         1      RC2 -0.36381959      3
12 ADMIN-UCSF-bo005     3         1      RC2  0.26820692      3
13 ADMIN-UCSF-bo008     3         1      RC2 -2.24123725      3
14 ADMIN-UCSF-bo010     3         1      RC2 -0.25274997      3
15 ADMIN-UCSF-bo011     3         1      RC2  0.31360819      3
16 ADMIN-UCSF-bo012     3         1      RC2 -0.03934867      3
17 ADMIN-UCSF-bo013     3         1      RC2 -0.06236109      3
18 ADMIN-UCSF-bo014     3         1      RC2  0.80369267      3
19 ADMIN-UCSF-bo015     3         1      RC2  0.41606599      3
20 ADMIN-UCSF-bo016     3         1      RC2 -0.06988054      3
21 ADMIN-UCSF-bo004     3         1      RC3 -2.69157047      3
22 ADMIN-UCSF-bo005     3         1      RC3 -0.82291746      3
23 ADMIN-UCSF-bo008     3         1      RC3 -1.52186068      3
24 ADMIN-UCSF-bo010     3         1      RC3 -0.59007055      3
25 ADMIN-UCSF-bo011     3         1      RC3  0.58379019      3
26 ADMIN-UCSF-bo012     3         1      RC3 -0.25388839      3
27 ADMIN-UCSF-bo013     3         1      RC3  1.22197350      3
28 ADMIN-UCSF-bo014     3         1      RC3 -1.63335701      3
29 ADMIN-UCSF-bo015     3         1      RC3  1.24595192      3
30 ADMIN-UCSF-bo016     3         1      RC3  0.01912759      3

Thanks. That's helpful as is the link. When I change it though, I get the error: ``` Error in eval(lhs, parent, parent) : argument "dat" is missing, with no default ```. I google that: it brings me here "https://stackoverflow.com/questions/24587341/once-again-setting-the-environment-within-a-function" which is a different situation then what I'm doing here. Would I not receive this error if I were using a tibble? — James, Apr 12 '20 at 20:46
Sorry, my bad. Editing it now: you need to specify `dat` when you call `cohort.fun` inside `melt.fun` — davidnortes, Apr 12 '20 at 20:55

Writing function...receiving error that variable isn't found

1 Answers1