CODEWITHSUNDEEP
rxts
Eka
Eka

Reputation: 15002

How to replace Inf and NaN with NA in an xts object

I have an xts object (df):

##                  a     b
## 2015-09-14 -0.5470    NA
## 2015-09-15  0.0112    NA
## 2015-09-16  0.0910 0.932
## 2015-09-17    -Inf 0.862
## 2015-09-18     Inf 1.946
## 2015-09-21 -0.7050 2.692
## 2015-09-22     NaN 2.011
## 2015-09-23 -0.5440 1.859
##     .
##     .
##    etc

Because column a includes Inf, -Inf and NaN, I am getting some errors. Is there a way to replace these values with NA (to treat Inf and NaN as missing values)?

Upvotes: 1

Views: 4697

Answers (1)

jbaums
jbaums

Reputation: 27408

You can use apply, like you might on an ordinary matrix

Using the data that you provided:

d <- xts(read.zoo(text='date a     b
2015-09-14 -0.5470    NA
2015-09-15  0.0112    NA
2015-09-16  0.0910 0.932
2015-09-17    -Inf 0.862
2015-09-18     Inf 1.946
2015-09-21 -0.7050 2.692
2015-09-22     NaN 2.011
2015-09-23 -0.5440 1.859', header=T, index=1))

apply(d, 2, function(x) ifelse(is.finite(x), x, NA))

##                  a     b
## 2015-09-14 -0.5470    NA
## 2015-09-15  0.0112    NA
## 2015-09-16  0.0910 0.932
## 2015-09-17      NA 0.862
## 2015-09-18      NA 1.946
## 2015-09-21 -0.7050 2.692
## 2015-09-22      NA 2.011
## 2015-09-23 -0.5440 1.859

An alternative approach is to make the replacement by indexing the elements to keep (or remove)

d[!is.finite(d)] <- NA

Upvotes: 4

Related Questions