I am trying to pass a variable Phyla (which is also the name of a df column of interest) into other functions. However I get the error: Error: Columntax_levelis unknown. Which I understand. It would just be more convenient to state the column you want to use once in the function since this will also be repeated numerous times in the script. I Have tried using OTU_melt_grouped[,1] since this will always be the first column to use in the dcast function, but get the error: Error: Must use a vector in[, not an object of class matrix. Moreover, it does not solve my solution in the group_by function since I want to be able to specify Phyla, Class, Order etc...
I am sure there must be a simple solution, but I don't know where to start!
taxa_specific_columns_func <- function(data, tax_level = Phyla) {
OTU_melt_grouped <- data %>%
group_by(tax_level, variable) %>%
summarise(value = sum(value))
taxa_cols <- dcast(OTU_melt_grouped, variable ~ tax_level)
rownames(taxa_cols) <- meta_data$site
taxa_cols <- taxa_cols[-1]
return(taxa_cols)
}
tax_test <- taxa_specific_columns_func(OTU_melt)
As we are passing an unquoted variable, we could make use of curly-curly ({{..}}) operator in group_by
library(dplyr)
library(tidyr)
library(tibble)
taxa_specific_columns_func <- function(data, tax_level = Phyla) {
data %>%
group_by({{tax_level}}, variable) %>%
summarise(value = sum(value)) %>%
pivot_wider(names_from = {{tax_level}}, values_from = value) %>%
column_to_rownames("variable")
}
taxa_specific_columns_func(OTU_melt)
# A B C D E
#a 0.01859254 0.42141238 -0.196961 -0.1859115 -0.2901680
#b -0.64700080 NA -0.161108 NA NA
#c -0.03297331 0.05871052 -1.963341 NA 0.7608218
data
set.seed(48)
OTU_melt <- data.frame(Phyla = rep(LETTERS[1:5], each = 3),
variable = sample(letters[1:3], 15, replace = TRUE), value = rnorm(15))
Related
My question seems simple, but I just can't do it. I have a dataframe with multiple columns with the name starting with coa and another column p with values like A, D, F, and so on, which changes according to the id.
All I found is how to do this matching with a fixed value, let's say "A", as below:
df <-df %>%
mutate(ly = any(str_detect(c_across(starts_with("coa")), "A")))
However, in my case, I want to compare to the column p specifically, where p changes, something like this:
df <-df %>%
mutate(ly = any(str_detect(c_across(starts_with("coa")), p)))
In this case, I get the error:
x no applicable method for 'type' applied to an object of class "factor"
Any thoughts? Thanks!
If we need to create a column, use if_any
library(dplyr)
library(stringr)
df <- df %>%
mutate(ly = if_any(starts_with("coa"), ~ str_detect(.x, p)))
I think this is a good place to use dplyr::across. You can run vignette('colwise') for a more comprehensive guide, but the key point here is that we can mutate all columns starting with "coa" simultaneously using the function == and we can pass a second argument, p, to == using the ... option provided by across.
library(dplyr)
df <- tibble(p = 1:10, coa1 = 1:10, coa2 = 11:20)
df %>%
mutate(across(.cols = starts_with('coa'), .fns = `==`, p))
This question already has answers here:
How to pass column name as argument to function for dplyr verbs?
(4 answers)
Closed 7 months ago.
I am trying to use group_by within a function call in dplyr (R) and I am getting unexpected results. Here is an example of what I am trying to do:
df = data.frame(a = c(0,0,1,1), b = c(0,1,0,1), c = c(1,2,3,4))
result1 = df %>%
group_by(a,b) %>%
mutate(d = sum(c))
result1$d
myFunc <- function(df, var) {
output = df %>%
group_by(a,!!var) %>%
mutate(d = sum(c))
return(output)
}
result2 = myFunc(df,"b")
result2$d
result1$d yields [1,2,3,4] which is what I expected. result2$d yields [3,3,7,7] which I do not want, and I am not sure what is going on.
It works to have b (without quotes) as the function argument, and {{var}} in place of !!var. Unfortunately, in my case, my column names are in string format (but maybe there is a way to transform the string beforehand so that it will work with the {{}} notation?)
If you want to pass a character object that can refer to a certain column of a data frame, you should use !!sym(var):
myFunc <- function(df, var) {
output = df %>%
group_by(a, !!sym(var)) %>%
mutate(d = sum(c))
return(output)
}
myFunc(df, "b")
If you want to pass a data-masked argument, you should use {{ var }} or equivalently !!enquo(var):
myFunc <- function(df, var) {
output = df %>%
group_by(a, {{ var }}) %>%
mutate(d = sum(c))
return(output)
}
myFunc(df, b)
Note that I pass "b" and b respectively into the function in the two different cases.
If we want to use quoting and unquoting instead of curlycurly {{}} the we should consider this basic procedure: https://tidyeval.tidyverse.org/dplyr.html
Creating a function around dplyr pipelines involves three steps: abstraction, quoting, and unquoting.
1. Abstraction step:
Here we identify the varying steps. In our case var in group_by:
2. Quoting step:
Identify all the arguments where the user is allowed to refer to data frame columns directly.
The function can’t evaluate these arguments right away.
Instead they should be automatically quoted. Apply enquo() to these arguments
3. Unquoting step:
Identify where these variables are passed to other quoting functions and unquote with !!.
In this case we pass var to group_by():
myFunc <- function(df, var) {
var <- enquo(var)
output = df %>%
group_by(a,!!var) %>%
mutate(d = sum(c))
return(output)
}
result2 = myFunc(df,b)
output:
[1] 1 2 3 4
Just as I post a question, I come across something that works...
myFunc <- function(df, var) {
output = df %>%
group_by_at(.vars = c("a",var)) %>%
mutate(d = sum(c))
return(output)
}
result2 = myFunc(df,"b")
Calculating a function of multiple variables for a dataframe in wide format is very familiar:
library(tidyverse)
df <- tibble(t = 1:3, b = 11:13, c = 21:23)
df <- df %>% mutate(d = b + c) # or base R: df$d <- df$b + df$c
What about when the dataframe is in long format? e.g.
df <- df %>% pivot_longer(-t, names_to = "variable", values_to = "value")
In this long format, you could imagine the same operation working by first group_by(t), and then calculating one value of d for each group, namely that group's variable=b value plus that group's variable=c value. Is this possible? One might think of something like summarise(d = b + c) but that expects wide format.
NB my real-world example has more than two cols b and c and I want to put them into a defined function, not just add them. My working solution is pivoting a huge dataframe from long to wide, calling my multivariable function to define a new column, then pivoting back to long.
Edit: to make the real world example explicit, I need to call a defined function that treats its arguments differently, unlike sum. For example
my.func <- function(b, c) { b^c }
How could the variable d be calculated by applying this function to the values of b and c associated with the same value of t?
We can just do sum instead of +
library(dplyr)
library(tidyr)
df %>%
group_by(t) %>%
summarise(d =sum(value[variable %in% c('b', 'c')]))
If it is to apply the my.func, we need to extract the value that correspond to 'b', 'c'
df %>%
group_by(t) %>%
mutate(new = my.func(value[variable == 'b'], value[variable == 'c']))
With the new release of dplyr I am refactoring quite a lot of code and removing functions that are now retired or deprecated. I had a function that is as follows:
processingAggregatedLoad <- function (df) {
defined <- ls()
passed <- names(as.list(match.call())[-1])
if (any(!defined %in% passed)) {
stop(paste("Missing values for the following arguments:", paste(setdiff(defined, passed), collapse=", ")))
}
df_isolated_load <- df %>% select(matches("snsr_val")) %>% mutate(global_demand = rowSums(.)) # we get isolated load
df_isolated_load_qlty <- df %>% select(matches("qlty_good_ind")) # we get isolated quality
df_isolated_load_qlty <- df_isolated_load_qlty %>% mutate_all(~ factor(.), colnames(df_isolated_load_qlty)) %>%
mutate_each(funs(as.numeric(.)), colnames(df_isolated_load_qlty)) # we convert the qlty to factors and then to numeric
df_isolated_load_qlty[df_isolated_load_qlty[]==1] <- 1 # 1 is bad
df_isolated_load_qlty[df_isolated_load_qlty[]==2] <- 0 # 0 is good we mask to calculate the global index quality
df_isolated_load_qlty <- df_isolated_load_qlty %>% mutate(global_quality = rowSums(.)) %>% select(global_quality)
df <- bind_cols(df, df_isolated_load, df_isolated_load_qlty)
return(df)
}
Basically the function does as follows:
1.The function selects all of the values of a pivoted dataframe and aggregated them.
2.The function selects the quality indicator (character) of a pivoted dataframe.
3.I convert the characters of the quality to factors and then to numeric to get the 2 levels (1 or 2).
4.I replace the numeric values of each of the individual columns by 0 or 1 depending on the level.
5.I rowsum the individual quality as I will get 0 if all of the values are good, otherwise the global quality is bad.
The problem is that I am getting the following messages:
1: `funs()` is deprecated as of dplyr 0.8.0.
Please use a list of either functions or lambdas:
# Simple named list:
list(mean = mean, median = median)
# Auto named with `tibble::lst()`:
tibble::lst(mean, median)
# Using lambdas
list(~ mean(., trim = .2), ~ median(., na.rm = TRUE))
This warning is displayed once every 8 hours.
Call `lifecycle::last_warnings()` to see where this warning was generated.
2: `mutate_each_()` is deprecated as of dplyr 0.7.0.
Please use `across()` instead.
I did multiple trials as for instance:
df_isolated_load_qlty %>% mutate(across(.fns = ~ as.factor(), .names = colnames(df_isolated_load_qlty)))
Error: Problem with `mutate()` input `..1`.
x All unnamed arguments must be length 1
ℹ Input `..1` is `across(.fns = ~as.factor(), .names = colnames(df_isolated_load_qlty))`.
But I am still a bit confused about the new dplyr syntax. Would someone be able to guide me a little bit around the right way of doing this?
mutate_each has been long deprecated and was replaced with mutate_all.
mutate_all is now replaced with across
across has default .cols as everything() which means it behaves as mutate_all by default (like here) if not mentioned explicitly.
You can apply the mulitple function in the same mutate call, so here factor and as.numeric can be applied together.
Considering all this you can change your existing function to :
library(dplyr)
processingAggregatedLoad <- function (df) {
defined <- ls()
passed <- names(as.list(match.call())[-1])
if (any(!defined %in% passed)) {
stop(paste("Missing values for the following arguments:",
paste(setdiff(defined, passed), collapse=", ")))
}
df_isolated_load <- df %>%
select(matches("snsr_val")) %>%
mutate(global_demand = rowSums(.))
df_isolated_load_qlty <- df %>% select(matches("qlty_good_ind"))
df_isolated_load_qlty <- df_isolated_load_qlty %>%
mutate(across(.fns = ~as.numeric(factor(.))))
df_isolated_load_qlty[df_isolated_load_qlty ==1] <- 1
df_isolated_load_qlty[df_isolated_load_qlty==2] <- 0
df_isolated_load_qlty <- df_isolated_load_qlty %>%
mutate(global_quality = rowSums(.)) %>%
select(global_quality)
df <- bind_cols(df, df_isolated_load, df_isolated_load_qlty)
return(df)
}
I am making my first baby steps with non standard evaluation (NSE) in dplyr.
Consider the following snippet: it takes a tibble, sorts it according to the values inside a column and replaces the n-k lower values with "Other".
See for instance:
library(dplyr)
df <- cars%>%as_tibble
k <- 3
df2 <- df %>%
arrange(desc(dist)) %>%
mutate(dist2 = factor(c(dist[1:k],
rep("Other", n() - k)),
levels = c(dist[1:k], "Other")))
What I would like is a function such that:
df2bis<-df %>% sort_keep(old_column, new_column, levels_to_keep)
produces the same result, where old_column column "dist" (the column I use to sort the data set), new_column (the column I generate) is "dist2" and levels_to_keep is "k" (number of values I explicitly retain).
I am getting lost in enquo, quo_name etc...
Any suggestion is appreciated.
You can do:
library(dplyr)
sort_keep=function(df,old_column, new_column, levels_to_keep){
old_column = enquo(old_column)
new_column = as.character(substitute(new_column))
df %>%
arrange(desc(!!old_column)) %>%
mutate(use = !!old_column,
!!new_column := factor(c(use[1:levels_to_keep],
rep("Other", n() - levels_to_keep)),
levels = c(use[1:levels_to_keep], "Other")),
use=NULL)
}
df%>%sort_keep(dist,dist2,3)
Something like this?
old_column = "dist"
new_column = "dist2"
levels_to_keep = 3
command = "df2bis<-df %>% sort_keep(old_column, new_column, levels_to_keep)"
command = gsub('old_column', old_column, command)
command = gsub('new_column', new_column, command)
command = gsub('levels_to_keep', levels_to_keep, command)
eval(parse(text=command))