CODEWITHSUNDEEP
rdataframe
Carbo
Carbo

Reputation: 916

erase rows when a variable is equal to 0 but not when it is NA , or when two columns are equal to NA in R

I have a dataset df that looks like this 

rel freq key
1.1 1.1 a       
1   1   b       
1   NA  c       
NA  NA  d       
NA  0   e       
NA  1   f

I want to obtain this:

rel freq key
1.1 1.1 a       
1   1   b       
1   NA  c               
NA  1   f

basically I want to get rid of the rows where freq = 0 or the rows where both rel and freq are equal to NA.

I tried 

df <- df[df$freq!=0,]

to filter out freq = 0 but it gives me some problems

similarly

library(dplyr)
df <- filter(df, freq != 0)

eliminates all the rows with freq = NA

Any suggestion on how I can solve this?

Upvotes: 0

Views: 46

Answers (2)

nurandi
nurandi

Reputation: 1618

df[!(!is.na(df$freq) & df$freq == 0 | (is.na(df$freq) & is.na(df$rel))),]

#   rel freq key
# 1 1.1  1.1   a
# 2 1.0  1.0   b
# 3 1.0   NA   c
# 6  NA  1.0   f

Data

df <- data.frame(rel = c(1.1, 1, 1, NA, NA, NA),
                 freq = c(1.1, 1, NA, NA, 0, 1),
                 key = letters[1:6])

Upvotes: 1

arg0naut91
arg0naut91

Reputation: 14764

Try:

subset(df, !(freq == 0 & !is.na(freq) | (is.na(freq) & is.na(rel))))

Output:

  rel freq key
1 1.1  1.1   a
2 1.0  1.0   b
3 1.0   NA   c
6  NA  1.0   f

Or in dplyr:

dplyr::filter(df, !(freq == 0 & !is.na(freq) | (is.na(freq) & is.na(rel))))

Upvotes: 2

Related Questions