Clean bad data automatically [duplicate]

Question

This question already has answers here:

How to convert a factor to integer\numeric without loss of information? (11 answers)

Closed 8 years ago.

I am building an App using shiny and openair to analyze wind data.
Right now the data needs to be “cleaned” before uploading by the user. I am interested in doing this automatically. Some of the data is empty, some of is not numeric, so it is not possible to build a wind rose. I want to:

1. Estimate how much of the data is not numeric
2. Cut it out and leave only numeric data

here is an example of the data:
the "NO2.mg" is read as a factor and not int becuse it does not consist only numbers
OK
here is a reproducible example:

no2<-factor(c(5,4,"c1",54,"c5",seq(2:50)))
no2
[1] 5  4  c1 54 c5 1  2  3  4  5  6  7  8  9  10 11 12 13 14
[20] 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
[39] 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
52 Levels: 1 10 11 12 13 14 15 16 17 18 19 2 20 21 22 ... c5
> as.numeric(no2)
[1] 45 34 51 46 52  1 12 23 34 45 47 48 49 50  2  3  4  5  6
[20]  7  8  9 10 11 13 14 15 16 17 18 19 20 21 22 24 25 26 27
[39] 28 29 30 31 32 33 35 36 37 38 39 40 41 42 43 44

Answer 1

To convert a factor to numeric, you need to convert to character first:

no2<-factor(c(5,4,"c1",54,"c5",seq(2:50)))
no2_num <- as.numeric(as.character(no2)) 
#Warning message:
#  NAs introduced by coercion 
no2_clean <- na.omit(no2_num) #remove NAs resulting from the bad data

# [1]  5  4 54  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36
# [40] 37 38 39 40 41 42 43 44 45 46 47 48 49
# attr(,"na.action")
# [1] 3 5
# attr(,"class")
# [1] "omit"

length(attr(no2_clean,"na.action"))/length(no2)*100
#[1] 3.703704

Answer 2

Worst R haiku ever:

Some of the data is empty, 
some of is not numeric, 
so it is not possible to build a wind rose.

Answer 3

OK this is how i did it i am sure someone has abetter way
i'd love it if you share with me
this is my data:
no2<-factor(c(5,4,"c1",54,"c5",seq(2:50)))
to count the "bad data:"

sum(is.na((as.numeric(as.vector(no2)))))

and to estimate the percent of bad data:
sum(is.na((as.numeric(as.vector(no2)))))/length(no2)*100