Merge and compare different columns from different files

问题

I'm trying to automate a process I've normally done in excel. This process consists of merge and compare different columns. For example:

df1: 
sp|P07437|TBB5_HUMAN
sp|P10809|CH60_HUMAN
sp|P424|LPPRC_HUMAN
sp|P474|LRC_HUMAN

df2: 
sp|P07437|TBB5_HUMAN
sp|P10809|CH60_HUMAN
sp|P42704|LPPRC_HUMAN

df3: 
sp|P07437|TBB5_HUMAN
sp|P10788|CH70_HUMAN
sp|P42704|LPPRC_HUMAN

And the output is something like that:

sp|P07437|TBB5_HUMAN  | sp|P07437|TBB5_HUMAN | sp|P07437|TBB5_HUMAN
sp|P10809|CH60_HUMAN  | sp|P10809|CH60_HUMAN |
                      |                      | sp|P10788|CH70_HUMAN
sp|P424|LPPRC_HUMAN   |                      |
sp|P474|LRC_HUMAN     |                      |
                      | sp|P42704|LPPRC_HUMAN| sp|P42704|LPPRC_HUMAN

I was trying to use the function compare or mergelink but I don't have this result. Do you know another function that I can use in this case?

More or less is something like Venn Diagram, that is exactly what I do after this in order to check that everything is good.

Here you are and a reproducible example:

df1 = data.frame(TEST1=c("sp|P07437|TBB5_HUMAN","sp|P10809|CH60_HUMAN", "sp|P424|LPPRC_HUMAN"))

df2 = data.frame(TEST2=c("sp|P07437|TBB5_HUMAN","sp|P10809|CH60_HUMAN","   sp|P42704|LPPRC_HUMAN"))

df3 = data.frame(TEST3=c("sp|P07437|TBB5_HUMAN","sp|P10788|CH70_HUMAN",     "sp|P42704|LPPRC_HUMAN"))

Thank you very much.

回答1:

I'm using a slightly-modified version of your data, avoiding factors in the data. I also trimmed extra white-space, assuming it's a mistake in copy/paste.

df1 = data.frame(TEST1=c("sp|P07437|TBB5_HUMAN","sp|P10809|CH60_HUMAN", "sp|P424|LPPRC_HUMAN"),
                 stringsAsFactors = FALSE)
df2 = data.frame(TEST2=c("sp|P07437|TBB5_HUMAN","sp|P10809|CH60_HUMAN","   sp|P42704|LPPRC_HUMAN"),
                 stringsAsFactors = FALSE)
df3 = data.frame(TEST3=c("sp|P07437|TBB5_HUMAN","sp|P10788|CH70_HUMAN",     "sp|P42704|LPPRC_HUMAN"),
                 stringsAsFactors = FALSE)

Since this kind of problem can easily extend to include more than the initial count of data.frames, I usually prefer to work with lists of data.frames, not explicit data.frames, if at all possible.

lst <- list(df1, df2, df3)

Now here's one method to get your desired results:

alltests <- unique(trimws(unlist(lst, recursive = TRUE)))
as.data.frame(
  setNames(lapply(lst, function(a) alltests[ match(alltests, a[,1]) ]),
           sapply(lst, names)),
  stringsAsFactors = FALSE
)
#                  TEST1                TEST2                TEST3
# 1 sp|P07437|TBB5_HUMAN sp|P07437|TBB5_HUMAN sp|P07437|TBB5_HUMAN
# 2 sp|P10809|CH60_HUMAN sp|P10809|CH60_HUMAN                 <NA>
# 3  sp|P424|LPPRC_HUMAN                 <NA>                 <NA>
# 4                 <NA>                 <NA>  sp|P424|LPPRC_HUMAN
# 5                 <NA>                 <NA> sp|P10809|CH60_HUMAN

This relies on (1) single-column data.frames (though that can be remedied); and (2) unique column names. Your suggested output did not imply any sort, so I opted to not do any sorting here; it's easy enough to use alltests <- sort(unique(...)), though note that it's an alphabetic sort, not based on the numeric portion of substrings.

来源：https://stackoverflow.com/questions/43266579/merge-and-compare-different-columns-from-different-files

标签

merge

multiple-columns