read_csv reads strings with "-" as date

[felixhaass]

read_csv() reads a string that contains a - always as a date, e.g. 3-19, but throws an error because the date is not complete. In the dataset I'm using, this denotes an ID string, however, and should be preserved as a string.

Here's a minimal working example:

fileConn<-file("test.csv")
writeLines(c("id, value","3-19, 5.0"), fileConn)
close(fileConn)

df <- read_csv("test.csv")

This produces:

> df <- read_csv("test.csv")
Warning message:
1 problems parsing 'test.csv'. See problems(...) for more details. 
> problems(df)
Source: local data frame [1 x 4]

  row col           expected actual
1   1   1 date like %Y-%m-%d   3-19

Now, this ends up as NA:

> df
Source: local data frame [1 x 2]

    id value
1 <NA>     5

[hadley]

More minimal reprex:

readr::read_csv("x,y
1989-90,1
1990-91,2")

Likely because CollectorDate::canParse() is accepting partial matches

[Huade]

Works if specify col_types as character:

readr::read_csv("period\n2008-2012\n2009-2013", col_types = "c")

[hadley]

Even more minimal reprex

collectorGuess(c("1989-90", "1990-91"))