Values change when using as.numeric() in r

Question

I've looked at other answers regarding my issue and nothing seems to help. So I hope I'm in the right to post a question similar to one that was posted some years ago.

My issue is that during read.table() all the columns come in as factors, even if some of them only contain values (and a few NA's), and, when I try to convert them using as.numeric(), the values simply change, although, indeed they turn numeric. See example.

df <- read.table(doc.csv, header = TRUE, sep = ",", dec = ".")
df$value

# Results in
[1]  2254    1873    2201    2147    2456    1785

# So..
as.numeric(df$value)
[1] 26 14 22 20 32 11

I have made sure that the original .csv document contains only numerics and NA's.
I have tried converting into characters first, and then numerics. Same issue.

Any ideas? I'm stumped.

Can you include a `dput(df)` or `str(df)`. Maybe add `stringsAsFactors = FALSE` to the read.table function and try again? — SeGa, Jun 01 '18 at 14:58
Short answer from the R-FAQ: use `as.numeric(as.character(x))` to convert a factor `x` to numeric. — Gregor Thomas, Jun 01 '18 at 15:00
Good point! As it turned out, dput() showed me that there were spaces in there with the values. Solved the issue when I removed them. Thanks for the tip. " 1178 ", " 1222 ", " 1223 ", " 1314 ", " 1462 ", — Þórgnýr Thoroddsen, Jun 01 '18 at 15:05
@Gregor - I know. I had tried that. But I found the issue as stated above. — Þórgnýr Thoroddsen, Jun 01 '18 at 15:06
Also. @Gregor. I can remove this question and add the solution to my specific issue to the other question. — Þórgnýr Thoroddsen, Jun 01 '18 at 15:08
Ah, for that just set `strip.white = TRUE` in your `read.table` call. — Gregor Thomas, Jun 01 '18 at 15:08
But also I'm not wholly convinced that is your problem, `as.numeric(" 123 ")` works just fine, as does `as.numeric(as.character(factor(" 123 ")))` — Gregor Thomas, Jun 01 '18 at 15:09
Good tip. Well, it fixed the issue for me. But of course the first resolution from here on will likely be the as.numeric(as.character(x)) — Þórgnýr Thoroddsen, Jun 01 '18 at 15:13

score 1 · Answer 1 · answered Jun 01 '18 at 15:00

1

you can try

as.numeric(as.character(df$value))

But I would say the reason that read.table reads it in as factors is because somewhere in your column there is non numeric data

answered Jun 01 '18 at 15:00

Shinobi_Atobe

1,793
1
18
35

Hi. I did try that before and it didn't work. But actually, I did solve the issue. A simple oversight on my behalf, there happened to be two spaces after the values in the csv document that I hadn't noticed. – Þórgnýr Thoroddsen Jun 01 '18 at 15:04

Values change when using as.numeric() in r

1 Answers1