Mistake in French locale

[briatte]

Hello,

There is a parsing issue in the French locale:

> dmy("3 janvier 2013", locale = "fr_FR")
[1] "2013-01-03 UTC"
> dmy("3 juillet 2013", locale = "fr_FR")
[1] NA
Warning message:
All formats failed to parse. No formats found.

The month juillet is the month july.

If this is not an issue with lubridate itself, please let me where shall I report this?

[vspinu]

Works fine for me under linux:

dmy("3 janvier 2013", locale = "fr_FR.utf8")
[1] "2013-01-03 UTC"

What do you get on

lubridate:::.build_locale_regs("fr_FR.utf8")$alpha_exact[["b"]]

A very long thread #181 might be relevant here, but would good to know more about the system.

[briatte]

You are right, the problem is on my end:

[1] "((?<b_b_e>Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)|(?<B_b_e>January|February|March|April|May|June|July|August|September|October|November|December))(?![[:alpha:]])"
Warning message:
In Sys.setlocale("LC_TIME", locale) :
  OS reports request to set locale to "fr_FR.utf8" cannot be honored

I am running Mac OS X, with the whole system set to English.

Session info:

> sessionInfo()
R version 3.0.1 (2013-05-16)
Platform: x86_64-apple-darwin10.8.0 (64-bit)

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

loaded via a namespace (and not attached):
 [1] colorspace_1.2-2   dichromat_2.0-0    digest_0.6.3       ggplot2_0.9.3.1    grid_3.0.1        
 [6] gtable_0.1.2       labeling_0.2       MASS_7.3-27        munsell_0.4.2      plyr_1.8          
[11] proto_0.3-10       RColorBrewer_1.0-5 reshape2_1.2.2     scales_0.2.3       stringr_0.6.2     
[16] tools_3.0.1       

It'd be useful to clarify why juillet is the only month that does not parse, though. The rest of the months work fine.

[hadley]

On OS X the locale is called fr_FR:

> lubridate:::.build_locale_regs("fr_FR")$alpha_exact[["b"]]
[1] "((?<b_b_e>jan|fév|mar|avr|mai|jui|jul|aoû|sep|oct|nov|déc)|(?<B_b_e>janvier|février|mars|avril|mai|juin|juillet|août|septembre|octobre|novembre|décembre))(?![[:alpha:]])"

It's not obvious why julliet fails to match that regexp though.

[briatte]

Same thing here:

> lubridate:::.build_locale_regs("fr_FR")$alpha_exact[["b"]]
[1] "((?<b_b_e>jan|fév|mar|avr|mai|jui|jul|aoû|sep|oct|nov|déc)|(?<B_b_e>janvier|février|mars|avril|mai|juin|juillet|août|septembre|octobre|novembre|décembre))(?![[:alpha:]])"

And it does work elsewhere:

> regexpr("janvier|février|mars|avril|mai|juin|juillet|août|septembre|octobre|novembre|décembre", "13 juillet 2013")
[1] 4
attr(,"match.length")
[1] 7
> grepl("janvier|février|mars|avril|mai|juin|juillet|août|septembre|octobre|novembre|décembre", "13 juillet 2013")
[1] TRUE

[vspinu]

I meant that juillet is parsed for me on linux:

dmy("3 juillet 2013", locale = "fr_FR.utf8")
[1] "2013-07-03 UTC"

So it is again a regexp issue on OS and it is getting closer to our Japanese friend problem.

[hadley]

Interestingly, it also fails in as.Date: http://stackoverflow.com/questions/17856415/mysterious-error-by-parsing-french-dates

[briatte]

R 3.0.3 has solved this issue:

• strptime() now checks the locale only when locale-specific formats are used and caches the locale in use: this can halve the time taken on OSes with slow system functions (e.g. OS X).
• strptime() and the format() methods for classes "POSIXct", "POSIXlt" and "Date" recognize strings with marked encodings: this allows, for example, UTF-8 French month names to be read on (French) Windows.

The mistake it caused in lubridate is gone:

library(lubridate)

clean_date = function(x) {

  # fix bugs in dates
  x = gsub("juillet", "07", x) # fix small bug in French month parser

  # parse to Date format
  x = parse_date_time(x, "%d %m* %Y", locale = "fr_FR.UTF-8")
  x = as.Date(x)

  return(x)
}

unclean_date = function(x) {

  # parse to Date format
  x = parse_date_time(x, "%d %m* %Y", locale = "fr_FR.UTF-8")
  x = as.Date(x)

  return(x)
}

clean_date("15 juillet 2007")
[1] "2007-07-15"

# returns NA on R < 3.0.3
unclean_date("15 juillet 2007")
[1] "2007-07-15"

The "juillet" month does not need fixing anymore.