Timestamp parsing error about nanotime

[eddelbuettel]

nanotime relies on the RcppCCTZ package which bundles CCTZ and the constraint seems to come from there:

R> library(RcppCCTZ
R> parseDatetime("2020-08-20 03:45:30.511400598", fmt ="%Y-%m-%d %H:%M:%E9S")
[1] "2020-08-20 03:45:30.5113 UTC"
R> parseDatetime("2020-0820 03:45:30.511400598", fmt ="%Y-%m%d %H:%M:%E9S")
[1] "2020-08-20 03:45:30.5113 UTC"
R> parseDatetime("20200820 03:45:30.511400598", fmt ="%Y%m%d %H:%M:%E9S")
Error in parseDatetime("20200820 03:45:30.511400598", fmt = "%Y%m%d %H:%M:%E9S") : 
  Parse error on 20200820 03:45:30.511400598
R> 

It's been a while since I looked but it may be documented behavior.

[eddelbuettel]

I just looked and there isn't too much there. Here is an example from time_tool and spacing is always used for the format.

[Famguy]

Ok. Thanks for your prompt response. I added a workaround by artificially inserting hypen into the strings and it works now.

[eddelbuettel]

Hi @Famguy

So as mentioned, nanotime relies on RcppCCTZ which wraps CCTZ. And I thought I had seen earlier issues to I just went back and checked. And indeed:

• google/cctz#137 recommends %E4Y for a four digit year (will test in a second)
• google/cctz#79 same
• google/cctz#25 same, also mentions %E4Y%M%D

all discuss it. We will make sure to fold some example use cases of %E4Y into our documentation. This should help you avoid having to insert hyphens.