R: Adding 3.1 and 0.2 resulting in a character column

Asked Feb 26 '18 at 16:38

Active Feb 26 '18 at 16:38

Viewed 21 times

I am doing a very simple operation where I create a new column in dataframe with values as 3.1+0.2. I want this new column as numeric and so I cast the operation as as.numeric. For some reason, the new column values are still characters, as can be seen below:

df <- data.frame(col=c(2, 3, 5))
df$newCol <- as.numeric(3.1+0.2)
filter(df, newCol == 3.3)

The filter for newCol == 3.3 results in zero rows.

[1] col    newCol
<0 rows> (or 0-length row.names)

However, if I filter newCol value as characters then it shows all records.

> filter(df, newCol == '3.3')
  col newCol
  1   2    3.3
  2   3    3.3
  3   5    3.3

It is weird that a simple operation converts the values into characters by default, and doesn't convert it into numeric even after using as.numeric.

asked Feb 26 '18 at 16:38

Gaurav

It's a floating point issue. Look at: `filter(df, round(newCol,1) == 3.3)` – arvi1000 Feb 26 '18 at 16:42
This is due to imprecision of storage of floating-point numbers. Do `options(digits=20); 3.3; 3.1+0.2`. – eipi10 Feb 26 '18 at 16:42
I think I understand the issue, but why does it convert into characters? – Gaurav Feb 26 '18 at 17:39
Neither column is being converted to character. You actully don't need the `as.numeric`. `newCol` will still be numeric. Run `str(df)` to see the class of each column. – eipi10 Feb 26 '18 at 17:59
It doesn‘t - check sapply(df, class) – RolandASc Feb 26 '18 at 18:00

R: Adding 3.1 and 0.2 resulting in a character column

0 Answers0