Error with ARIMA

Question

I'm trying to run an ARIMA on a temporal dataset that is in a .csv file. Here is my code so far:

Oil_all <- read.delim("/Users/Jkels/Documents/Introduction to Computational
Statistics/Oil production.csv",sep="\t",header=TRUE,stringsAsFactors=FALSE)
Oil_all

The file looks like:

year.mbbl
1     1880,30
2     1890,77
3    1900,149
4    1905,215
5    1910,328
6    1915,432
7    1920,689
8   1925,1069
9   1930,1412
10  1935,1655
11  1940,2150
12  1945,2595
13  1950,3803
14  1955,5626
15  1960,7674
16  1962,8882
17 1964,10310
18 1966,12016
19 1968,14104
20 1970,16690
21 1972,18584
22 1974,20389
23 1976,20188
24 1978,21922
25 1980,21732
26 1982,19403
27 1984,19608

Code:

apply(Oil_all,1,function(x) sum(is.na(x)))

Results:

[1] 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

When I run ARIMA:

library(forecast)
auto.arima(Oil_all,xreg=year)

This is the error:

Error in lm.fit(x, y, offset = offset, singular.ok = singular.ok, ...) : 
  0 (non-NA) cases
In addition: Warning message:
In data.matrix(data) : NAs introduced by coercion

So, I was able to call in the data set and it prints. However, when I go to check whether the values are present with the apply function, I see all 0's, so I know something's wrong and that's probably why I'm getting the error. I'm just not sure what the error means or how to fix it in the code.

Any advice?

Answer 1

If I got your question right, it should be like:

Oil_all <- read.csv("myfolder/myfile.csv",header=TRUE) 
## I don't have your source data, so I tried to reproduce it with the data you printed
Oil_all
   year value
1  1880    30
2  1890    77
3  1900   149
4  1905   215
5  1910   328
6  1915   432
7  1920   689
8  1925  1069
9  1930  1412
10 1935  1655
11 1940  2150
12 1945  2595
13 1950  3803
14 1955  5626
15 1960  7674
16 1962  8882
17 1964 10310
18 1966 12016
19 1968 14104
20 1970 16690
21 1972 18584
22 1974 20389
23 1976 20188
24 1978 21922
25 1980 21732
26 1982 19403
27 1984 19608


library(forecast)
auto.arima(Oil_all$value,xreg=Oil_all$year)

Series: Oil_all$value 
ARIMA(3,0,0) with non-zero mean 

Coefficients:
         ar1     ar2      ar3  intercept  Oil_all$year
      1.2877  0.0902  -0.4619  -271708.4      144.2727
s.e.  0.1972  0.3897   0.2275   107344.4       55.2108

sigma^2 estimated as 642315:  log likelihood=-221.07
AIC=454.15   AICc=458.35   BIC=461.92

Answer 2

your import should be

Oil_all<-read.csv("/Users/Jkels/Documents/Introduction to Computational Statistics/Oil production.csv")

That is why your data is weird. Sorry I do not have the reputation to comment.I did the same as Nemesi and it worked then. I think you are trying to import a csv as a tab delimited file.