How do I convert a 2x2 contingency table into a long format dataframe?

Question

How do I convert a 2x2 contingency table into a long format data frame? I tried this:

library(reshape2)
Table <- matrix(c(7,67,19,71), 2, 2, byrow=TRUE)
rownames(Table) <- c('Drug', 'No_Drug')
colnames(Table) <- c('Comp', 'No_Comp')
melt(Table)

I get this rather than a data frame of 164 rows categorized by Drug vs. No_Drug

  Var1    Var2 value
1    Drug    Comp     7
2 No_Drug    Comp    19
3    Drug No_Comp    67
4 No_Drug No_Comp    71

`data.frame(expand.grid(rownames(Table), colnames(Table)), value = c(Table))`? — MichaelChirico, Nov 07 '15 at 14:59

Steven Beaupré · Accepted Answer · 2015-11-07T15:23:40.560

1

From what I understand, you are trying to convert your 4x3 molten data.frame into a 164x2 data.frame. You could try expandRows() from the splitstackshape package:

Expands (replicates) the rows of a data.frame or a data.table, either by a fixed number, a specified vector, or a value contained in one of the columns in the source data.frame or a data.table.

In your case, simply do:

library(splitstackshape)
m <- melt(Table)
expandRows(m, "value")

Note: This is a convenient wrapper around base R:

out <- m[rep(sequence(nrow(m)), m[["value"]]), ]
out[["value"]] <- NULL

edited Nov 07 '15 at 15:23

answered Nov 07 '15 at 15:01

Steven Beaupré

21,343
7
57
77

I suggest that the moderators add a tag for the package splitstackshape – FTF Nov 07 '15 at 15:18
@FTF I added the tag. If you ever happen to have a question related to this package, you'll be able to use it. – Steven Beaupré Nov 07 '15 at 15:49
Thanks for your comment @Steven Beaupré. Might not library(dplyr) filter(out) ##work too? – FTF Nov 07 '15 at 16:30
It would be `dplyr::select(out, -value)` – Steven Beaupré Nov 07 '15 at 16:31

How do I convert a 2x2 contingency table into a long format dataframe?

1 Answers1