0

I have been forming a panel dataset based on many other panel datasets. I have been puzzled by data.table or even base R function merge() that often alter the ordering of rows. So after much trial and error I used the following from plyr:

join(x, y, by=c("Firm", "Index"))

Finally, when I prepared by final panel dataset and used the attach() function and plm package for the analysis, I get the following error:

Error in pdim.default(index[[1]], index[[2]]) when I investigate further using 

options(error=recover)

I find

stop(cat("duplicate couples (time-id)\n"))

Kindly suggest what could be wrong and how to fix it?

Gregor Thomas
  • 136,190
  • 20
  • 167
  • 294
Polar Bear
  • 731
  • 1
  • 7
  • 21
  • try `sort = FALSE` inside `merge()`. – mtoto May 19 '16 at 18:17
  • @Gregor why don't attach()? – Polar Bear May 20 '16 at 04:23
  • It can cause problems that are difficult to debug when attached columns get out of sync with each other (and with the data frame). `with()` is a good alternative for single commands that don't have a data argument. But many commands (like `plm`!) have a `data` argument so that you don't need to `attach` or use `with` at all. – Gregor Thomas May 20 '16 at 05:20
  • See the [answers here](http://stackoverflow.com/q/1310247/903061) for more details on the badness of `attach`. – Gregor Thomas May 20 '16 at 05:24

1 Answers1

2

It seems like in your data (maybe due to the merging process) you have individuals which have the same value in the time index more than once (or more than one NA). You could either look at your data or try table(index(your_pdataframe), useNA = "ifany") to find out which.

Helix123
  • 3,502
  • 2
  • 16
  • 36
  • Yes I intentionally have repeating values in Index. The provided code gives: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 . What does it mean? – Polar Bear May 20 '16 at 04:28