summary fails on a nanotime vector

[lsilvest]

summary(nanotime((1:100)*3600e9))
                                 Min.                               1st Qu. 
"1970-01-01T00:00:00.000000000+00:00" "1970-01-01T00:00:00.000000000+00:00" 
                               Median                                  Mean 
"1970-01-01T00:00:00.000000000+00:00" "1970-01-01T00:00:00.000000000+00:00" 
                              3rd Qu.                                  Max. 
"1970-01-01T00:00:00.000000000+00:00" "1970-01-01T00:00:00.000000000+00:00" 
Warning message:
In qtile.integer64(x, probs = probs, na.rm = na.rm, names = names,  :
  class 'nanotime' has no 'names' slot; assigning a names attribute will create an invalid object

(Edited: You want three opening ticks followed by the language, here r. --Dirk)

[eddelbuettel]

Not nice. I guess we need a new S4 method?

[lsilvest]

Yes, will look into it.

[eddelbuettel]

After #25 it does it as int64, but not always:

R> options("width"=70)
R> summary(nanotime(1:10))
                                 Min. 
"1970-01-01T00:00:00.000000001+00:00" 
                              1st Qu. 
"1970-01-01T00:00:00.000000003+00:00" 
                               Median 
"1970-01-01T00:00:00.000000006+00:00" 
                                 Mean 
"1970-01-01T00:00:00.000000005+00:00" 
                              3rd Qu. 
"1970-01-01T00:00:00.000000008+00:00" 
                                 Max. 
"1970-01-01T00:00:00.000000010+00:00" 
R> summary(nanotime(1:10)*3.6e6)
integer64
    Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
 3600000 10800000 21600000 19800000 28800000 36000000 
R> 

[lsilvest]

What's going on above is that multiplying a nanotime by a double reverts to an integer64. The following works as intended:

> summary(nanotime(1:10*3.6e6))
                                 Min.                               1st Qu. 
"1970-01-01T00:00:00.003600000+00:00" "1970-01-01T00:00:00.010800000+00:00" 
                               Median                                  Mean 
"1970-01-01T00:00:00.021600000+00:00" "1970-01-01T00:00:00.019800000+00:00" 
                              3rd Qu.                                  Max. 
"1970-01-01T00:00:00.028800000+00:00" "1970-01-01T00:00:00.036000000+00:00" 

I would say we should raise an error on multiplication and division of a nanotime, like POSIXct does:

> Sys.time()*3
Error in Ops.POSIXt(as.POSIXct(Sys.time()), 3) : 
  '*' not defined for "POSIXt" objects

What do you think?

[eddelbuettel]

That's probably reasonable, and doable.

[eddelbuettel]

Nicely done:

R> summary(nanotime((1:100)*3600e9))
                                 Min.                               1st Qu. 
"1970-01-01T01:00:00.000000000+00:00" "1970-01-02T02:00:00.000000000+00:00" 
                               Median                                  Mean 
"1970-01-03T02:00:00.000000000+00:00" "1970-01-03T02:30:00.000000000+00:00" 
                              3rd Qu.                                  Max. 
"1970-01-04T03:00:00.000000000+00:00" "1970-01-05T04:00:00.000000000+00:00" 
R>