compare two similar data frames and only display differences in R - r

Question

I have two similar data frames with the same number of columns, but different numbers of rows. Most of the entries between the two are the same, but in a few places there are differences, and these are what I care about. The first column in both data frames serves as a key.
Ideally, I'd like to be able to see whether they've changed, as well as the values from each of the two data frames. My first solution was to create a merged dataframe and re-organize the columns side-by-side like so:
df1<-data.frame(gene=c('cyp1a1','cyp2a6','srd5a','slc5a5','cox15'), updated=c(TRUE,TRUE,FALSE,TRUE,FALSE),version=c(2,3,1,2,1))
df2<-data.frame(gene=c('cyp1a1','cyp2a6','srd5a','slc5a5'), updated=c(FALSE,TRUE,FALSE,FALSE),version=c(1,2,1,1))
#merge data frames
comp<-merge(df1,df2, by="gene", all=TRUE)
#re-order columns side-by-side
#probably a better way to do this
ordList<-c(1,2,4,3,5)
comp<-comp[ordList]
So now I have a side-by-side comparison data frame. I am unsure about how to iterate over the data frames to perform the comparison. Eventually I would like to create a new data-frame which uses information from the comparison to exclude data that is identical (replace with empty string) and includes data that differs from the first df to the second.
This is what comp looks like now:
gene updated.x updated.y version.x version.y
1 cox15 FALSE NA 1 NA
2 cyp1a1 TRUE FALSE 2 1
3 cyp2a6 TRUE TRUE 3 2
4 slc5a5 TRUE FALSE 2 1
5 srd5a FALSE FALSE 1 1
This is what I want it to look like:
gene updated.x updated.y version.x version.y
1 cox15 FALSE NA 1 NA
2 cyp1a1 TRUE FALSE 2 1
3 cyp2a6 3 2
4 slc5a5 TRUE FALSE 2 1
5 srd5a
In my actual data, I have 14 columns in each data frame, and hundreds of rows. I may be doing similar comparisons in the future, so having a functional way of executing this task would be ideal.

Answer 1

Here is my suggestion, considering you have 14 columns:
library(data.table)
library(magrittr)
z = rbindlist(list(df1,df2), idcol=TRUE)
z[, lapply(.SD, . %>% unique %>% paste(collapse=";")), keyby=gene]
# gene .id updated version
# 1: cox15 1 FALSE 1
# 2: cyp1a1 1;2 TRUE;FALSE 2;1
# 3: cyp2a6 1;2 TRUE 3;2
# 4: slc5a5 1;2 TRUE;FALSE 2;1
# 5: srd5a 1;2 FALSE 1
This shows you which data frame each gene appears in (.id) as well as the attributes (updated and version). This display extends naturally to additional tables, like list(df1,df2,df3).
If you really are not interested in unchanged values, you can hide them with an if test:
z[, lapply(.SD, function(x)
if (uniqueN(x)>1) x %>% unique %>% paste(collapse=";")
else ""
), keyby=gene]
# gene .id updated version
# 1: cox15
# 2: cyp1a1 1;2 TRUE;FALSE 2;1
# 3: cyp2a6 1;2 3;2
# 4: slc5a5 1;2 TRUE;FALSE 2;1
# 5: srd5a 1;2
This also hides .id for genes only showing up once, but that can be tweaked.
Explanation. z contains all the data, "stacked" or stored in "long" format.
To make the summary table, we use z[, j, keyby=gene] where j works on the Subset of Data, .SD, associated with each keyby=gene group and returns a list of column vectors for the result.
The . %>% unique %>% paste(collapse=";") uses a feature of magrittr. It is just an easy-to-read version of function(y) paste(unique(y), collapse=";"). When it starts with x, it applies the function to x. You can replace it if you prefer to write these in the standard way.