format a time series as dataframe with julian date

Question

I have a time series tt.txt of daily data from 1st May 1998 to 31 October 2012 in one column as this:

Each number in the text file represents the maximum temperature in kelvin for the corresponding day. I want to put the data in 3 columns as follows by adding year, jday, and the value of the data:

     year jday MAX_TEMP 
1    1959  325 11.7      
2    1959  326 15.6      
3    1959  327 14.4

akrun · Accepted Answer · 2015-06-08T14:27:58.663

1

If you have a vector with dates, we can convert it to 'year' and 'jday' by

v1 <- c('May 1998 05', 'October 2012 10')
v2 <- format(as.Date(v1, '%b %Y %d'), '%Y %j')
df1 <- read.table(text=v2, header=FALSE, col.names=c('year', 'jday'))
df1
#  year jday
#1 1998  125
#2 2012  284

To convert back from '%Y %j' to 'Date' class

df1$date <- as.Date(do.call(paste, df1[1:2]), '%Y %j')

Update

We can read the dataset with read.table. Create a sequence of dates using seq if we know the start and end dates, cbind with the original dataset after changing the format of 'date' to 'year' and 'julian day'.

dat <- read.table('tt.txt', header=TRUE)
date <- seq(as.Date('1998-05-01'), as.Date('2012-10-31'), by='day')
dat2 <- cbind(read.table(text=format(date, '%Y %j'), 
              col.names=c('year', 'jday')),MAX_TEMP=dat[1])

edited Jun 08 '15 at 14:27

answered Jun 08 '15 at 13:24

akrun

874,273
37
540
662

@user3290596 Assuming that the date starts from '1998-05-01' and ends with '2012-10-31', we can create a sequence with `seq`, change the `format` and cbind with the initial dataset – akrun Jun 08 '15 at 14:29
an error occurred in dat2: dat2 <- cbind(read.table(text=format(date, '%Y %j'),col.names=c('year', 'jday')),MAX_TEMP=dat[1]) Error in data.frame(..., check.names = FALSE) : arguments imply differing number of rows: 5298, 5449 looks like length mismatch – user3290596 Jun 08 '15 at 18:43
@user3290596 The `seq` created is of a particular length. If your data i.e. Max_temp didn't have that many elements, it will cause that error. Without providing much details, it is hard to understand which day you have a missing value for MAX_TEMP – akrun Jun 08 '15 at 18:47

score 0 · Answer 2 · answered Jun 08 '15 at 13:25

0

You can use yday

as.POSIXlt("8 Jun 15", format = "%d %b %y")$yday

answered Jun 08 '15 at 13:25

jalapic

13,792
8
57
87

format a time series as dataframe with julian date

2 Answers2

Update