Somewhat cryptic error message/behavior with rbind_all

[kevinykuo]

When pulling data from different sources sometimes I have situations like the following. Is this expected behavior and should we have more informative error messages?

df1 <- data.frame(ID = c("A", "B"), value = as.integer(c(1, 2)))
df2 <- data.frame(ID = c("C", "D"), value = c("00003", "00004"))
dfList <- list(df1, df2)
do.call(rbind, dfList) # works fine
rbind_all(dfList) # Error: not compatible with STRSXP

[hadley]

This is expected behaviour, although I admit the error message could be better. It's basically telling you that rbind_list() won't automatically coerce an integer into a string, because that probably represents a problem with the underlying data.

[romainfrancois]

The problem was with combining integers with factors, which was erroneously allowed. We now the usual error message:

Erreur : incompatible type (data index: 2, column: 'value', was collecting: integer (dplyr::Collecter_Impl<13>), incompatible with data of type: factor
De plus : Message d'avis :
In rbind_all(dfList) : Unequal factor levels: coercing to character