[datatable-help] What does "problem re-suffixing the mergeddata.table" mean?

[Matthew Dowle]

Hi Allan,

I'm not sure about that as I don't use merge(),  Steve or Tom will know.

I use X[Y,j] syntax together with join inherited scope, as per FAQ 1.11
and 1.12. That is more efficient than merge.data.table (at least, it should 
be).

I think the reason the merge.data.table method exists in data.table is to 
make the
transition easier for users who are used to merge on data.frame.  However, 
the
idiom of merge() which is to create a large object with many columns, 
followed
by a calculation on the result, seems inefficient to me. See nuances in FAQ 
1.12.

I suspect I may be missing something, though. Do you see what I mean about
X[Y,j] and does it work for you?   If there is a good reason to use merge(), 
other
than people are used to it, I'll need a worked example to understand please.
Perhaps it is related to flexibility and not being tied to joining in the 
key columns?
If that's the case we should add whatever the advantage of merge() is into 
FAQs
1.11 and 1.12, at least.

Matthew


"Allan Engelhardt (CYBAEA)" <allane at cybaea.com> wrote in message 
news:4E3C0495.6060304 at cybaea.com...
> What does the warning mean?  What sort of problem?
>
> model.data <- merge(Y1, Y2, all = FALSE, suffixes = c(".Y1", ".Y2"))
> Warning message:
> In merge.data.table(Y1, Y2, all = FALSE, suffixes = c(".Y1", ".Y2")) :
>   There was a problem re-suffixing the merged data.table. The merge was 
> succesful, but the default column naming has been used (ie. suffixes were 
> set to c('', '.1').
>
> It doesn't seem to matter what value I use for suffixes=
>
> Allan
>
> -- 
> ---
> http://www.cybaea.net/Blogs/Data/