Number values include comma -- how do I make these numeric?

Question

I have a whole column of numbers that include comma separators at the thousands. When I try to create a numeric column out of them, anything over 999 becomes NA.

I used cbind:

df <- cbind(df, var2 = as.numeric(as.character(df$var1)))

and wound up with:

        var1  var2
1   2,518.50    NA
2   2,518.50    NA
3   5,018.50    NA
4   4,018.50    NA
5  10,018.50    NA
6     318.50 318.5
7   2,518.50    NA
8   3,518.50    NA
9   7,518.50    NA
10  1,018.50    NA

Is there a way to strip the commas or tell as.numeric how to handle them?

@mnel I think that would do something different? Wouldn't that treat `1,000` the same as `1`? Whereas Amanda is saying that `1,000` should be treated like `1000` — Dason, Nov 27 '12 at 22:27
@Justin, that works actually. I was too focused on telling as.numeric about formats to expect. — Amanda, Nov 27 '12 at 22:44
@Amanda, how did you get the original data into R? You may also want to google `stringsAsFactors=FALSE` — Ricardo Saporta, Nov 27 '12 at 22:48
parse_number() from the readr package does this well. It ignores the commas inside an number. — Magnus, May 28 '18 at 16:41

Ricardo Saporta · Accepted Answer · 2012-11-27T22:45:58.990

16

If you are trying to add a new column var2 to df, you can use the following

  df$var2 <- as.numeric(gsub(",", "", as.character(df$var1)))

edited Nov 27 '12 at 22:45

answered Nov 27 '12 at 22:36

Ricardo Saporta

54,400
17
144
178

1

Nice. I don't even need cbind? – Amanda Nov 27 '12 at 22:46
3

nope, with data frames you can just assign. `df[[var2]] <- value` and `df$var2 <- value` both work. – Ricardo Saporta Nov 27 '12 at 22:50

score 5 · Answer 2 · answered Nov 27 '12 at 22:37

5

Use as.numeric(gsub(",", "", df$var1)).

You want to use gsub as sub will only replace the first comma.

answered Nov 27 '12 at 22:37

Erik Shilts

4,389
2
26
51

Number values include comma -- how do I make these numeric?

2 Answers2

Linked