Delete missing values detected by ggplot() in R

Question

I asked this question to know how it is possible to plot many graphs in the same plot. Following to the answer which I liked and accepted, it is possible to use ggplot() function.

Now using ggplot(), I receive the following message which notifies that there are missing values were deleted for the plot:

Warning message:
Removed 33 row(s) containing missing values (geom_path).

From the produced plot and visualizing, I am satisfied with data after that ggplot() removed the 33 rows.

I know how to delete rows of NA but here I don't understand if ggplot() deleted rows where there exist NA for at least one variable OR removed rows where all variables are NA, knowing that I have 7 variables and there are some rows where all variables are completely NA while many rows contain NA for only some variables.

Question: Although the rows are already deleted for the plot, how it is possible to remove these rows "the detected 33 rows" completely from data?

zx8754 · Accepted Answer

ggplot removes rows with NA for columns that are used as input aes to ggplot, if input is x and y columns, but dataframe has y column as well, it will only drop rows if x or y has NA.

Here is an example:

library(ggplot2)

x <- head(mtcars)

# add NA to some column we don't use for ggplot
x$am[ 1 ] <- NA

ggplot(x, aes(cyl, mpg)) + geom_point()
# no warnings

# now add NA to column that we use for plotting
x$cyl[ 1 ] <- NA

ggplot(x, aes(cyl, mpg)) + geom_point()
# Warning message:
#   Removed 1 rows containing missing values (geom_point). 

# to avoid that warning, we can explicitly set it to remove NA
ggplot(x, aes(cyl, mpg)) + geom_point(na.rm = TRUE)
# no warnings

To remove rows from the data, check if the selected columns have NA:

x_clean <- x[ !(is.na(x$cyl) | is.na(x$mpg)), ]
ggplot(x_clean , aes(cyl, mpg)) + geom_point()
# no warnings

Edit 1: To apply to your data based on comments, try below, see filter:

Data <- bind_rows(...)
Data %>%
  mutate(data = paste0('Data',data)) %>%
  pivot_longer(-c(data,Time)) %>%
  filter(!(is.na(Time) | is.na(value))) %>% 
  ggplot(aes(x = factor(Time), y =value), group = name, color = name))+
  geom_line()+
  facet_wrap(.~data,scales = 'free', ncol = 1) +
  xlab('Time')

Edit 2: To "know" what data is going into ggplot why not keep filtered clean data as a separate object instead of piping, see:

Data <- bind_rows(...)
cleanData <- Data %>% 
  mutate(data = paste0('Data',data)) %>%
  pivot_longer(-c(data,Time)) %>%
  filter(!(is.na(Time) | is.na(value)))
  
ggplot(cleanData, aes(x = factor(Time), y =value), group = name, color = name)+
  geom_line()+
  facet_wrap(.~data,scales = 'free', ncol = 1) +
  xlab('Time')

Delete missing values detected by ggplot() in R

Answers (2)

Related Questions