Skip to content

write.table (PR#11625)

1 message · jdowty at unimelb.edu.au

#
Hi,=20

=20

I am writing to suggest a change in write.table().  Currently one
feature of this function is so counter-intuitive that it could be
considered to be a bug.  I use R 2.6.1 under Windows XP.=20=20

=20

When appending to a file, write.table() automatically writes the column
names first, even though this means writing the variable names in the
middle of a column of numeric values.  The function gives a warning but
if you have the function in a loop then usually all you will see is a
non-specific message ("there were 50 or more warnings") which can be
easy to miss.  In almost all cases it would make more sense for
"col.names=3DTRUE" to be interpreted to mean that the function should
print the column names if it has created the file but not otherwise, I
think.=20=20

=20

James.=20=20

=20

=20

=20