Combine two lists with different structure in R

Question

I have vendor_list

vendor_list[57:59]
[[1]]
[1] "ibm"

[[2]]
[1] "apache"    "canonical" "apple"     "novell"   

[[3]]
[1] "gnu"    "oracle"

And I have problemtype_list

problemtype_list[57:59]
[[1]]
[1] "NVD-CWE-Other"

[[2]]
[1] "NVD-CWE-Other"

[[3]]
[1] "CWE-824"

I need to combine them to make a data frame such that such that

A              B
ibm       NVD-CWE-Other
apache    NVD-CWE-Other
canonical NVD-CWE-Other
apple     NVD-CWE-Other
novelle   NVD-CWE-Other 
gnu       CWE-824
oracle    CWE-824

I have seen similar question Combine two lists in a dataframe in R

But it gives me error

do.call(rbind, Map(data.frame, A=problemtype_list, B=vendor_list))
Error in (function (..., row.names = NULL, check.rows = FALSE, check.names = TRUE,  : 
  arguments imply differing number of rows: 1, 0

EDIT

My structure of each list

str(vendor_list)
 $ : chr "cisco"
 $ : NULL
 $ : chr [1:5] "redhat" "novell" "debian" "oracle" ...
 $ : chr [1:4] "redhat" "novell" "debian" "google"
 $ : chr [1:4] "redhat" "novell" "debian" "google"

 str(problemtype_list)
 $ : chr "CWE-254"
 $ : chr "CWE-79"
 $ : chr "NVD-CWE-Other"
 $ : chr "NVD-CWE-Other"
 $ : chr "CWE-254"
 $ : chr "CWE-189"
 $ : chr "CWE-119"

(I was right ... you have an empty element in `vendor_list`!) — r2evans, Sep 29 '17 at 23:04
Actually empty elements would not throw an error. NULL, however, will. You can do a quick lapply to replace those NULL with empty strings or whatever you prefer. — Troy, Sep 29 '17 at 23:07

score 2 · Accepted Answer · answered Sep 29 '17 at 23:02

My guess is that one of your lists has a zero-length element.

vendor_list <- list("ibm", c("apache", "canonical", "apple", "novell"), c("gnu", "oracle"))
problemtype_list <- list("NVD-CWE-Other", "NVD-CWE-Other", "CWE-824")
do.call(rbind.data.frame, Map(data.frame, A=vendor_list, B=problemtype_list))
#           A             B
# 1       ibm NVD-CWE-Other
# 2    apache NVD-CWE-Other
# 3 canonical NVD-CWE-Other
# 4     apple NVD-CWE-Other
# 5    novell NVD-CWE-Other
# 6       gnu       CWE-824
# 7    oracle       CWE-824

However, if we provide an empty slot:

vendor_list[[3]] <- character(0)
vendor_list
# [[1]]
# [1] "ibm"
# [[2]]
# [1] "apache"    "canonical" "apple"     "novell"   
# [[3]]
# character(0)

... and a quick test:

any(lengths(vendor_list) == 0)
# [1] TRUE
any(lengths(problemtype_list) == 0)
# [1] FALSE

... then the merge fails:

do.call(rbind.data.frame, Map(data.frame, A=vendor_list, B=problemtype_list))
# Error in (function (..., row.names = NULL, check.rows = FALSE, check.names = TRUE,  (from pit-roads.R!8460QVH#21) : 
#   arguments imply differing number of rows: 0, 1

You can either replace the offending entries with something that makes sense (e.g., NA), or you can remove them. Which method you use depends entirely on your use.

Replacement:

vendor_list[lengths(vendor_list) == 0] <- NA
problemtype_list[lengths(problemtype_list) == 0] <- NA
do.call(rbind.data.frame, Map(data.frame, A=vendor_list, B=problemtype_list))
#           A             B
# 1       ibm NVD-CWE-Other
# 2    apache NVD-CWE-Other
# 3 canonical NVD-CWE-Other
# 4     apple NVD-CWE-Other
# 5    novell NVD-CWE-Other
# 6      <NA>       CWE-824

Removal:

keepthese <- (lengths(vendor_list) > 0) & (lengths(problemtype_list) > 0)
keepthese
# [1]  TRUE  TRUE FALSE
vendor_list <- vendor_list[keepthese]
problemtype_list <- problemtype_list[keepthese]
do.call(rbind.data.frame, Map(data.frame, A=vendor_list, B=problemtype_list))
#           A             B
# 1       ibm NVD-CWE-Other
# 2    apache NVD-CWE-Other
# 3 canonical NVD-CWE-Other
# 4     apple NVD-CWE-Other
# 5    novell NVD-CWE-Other

another alt `stack(setNames(vendor_list, problemtype_list))` — user20650, Sep 29 '17 at 23:12
I knew there was a simpler function out there ... thanks, @user20650! — r2evans, Sep 29 '17 at 23:19

score 1 · Answer 2 · answered Sep 29 '17 at 22:54

The code you say doesn't work works for me - I've called things p and v:

> v = list("ibm",c("apache","canonical","apple","novelle"),c("gnu","oracle"))

> p = list("NVD-CWE-Other","NVD-CWE-Other","CWE-824")

> p
[[1]]
[1] "NVD-CWE-Other"

[[2]]
[1] "NVD-CWE-Other"

[[3]]
[1] "CWE-824"

> v
[[1]]
[1] "ibm"

[[2]]
[1] "apache"    "canonical" "apple"     "novelle"  

[[3]]
[1] "gnu"    "oracle"

Then your code:

> do.call(rbind, Map(data.frame, A=p, B=v))
              A         B
1 NVD-CWE-Other       ibm
2 NVD-CWE-Other    apache
3 NVD-CWE-Other canonical
4 NVD-CWE-Other     apple
5 NVD-CWE-Other   novelle
6       CWE-824       gnu
7       CWE-824    oracle

So maybe your data is structured differently.

Alternatively:

> do.call(rbind.data.frame,mapply(cbind,v,p))
         V1            V2
1       ibm NVD-CWE-Other
2    apache NVD-CWE-Other
3 canonical NVD-CWE-Other
4     apple NVD-CWE-Other
5   novelle NVD-CWE-Other
6       gnu       CWE-824
7    oracle       CWE-824
>

I will add str output of each list. because I am still getting error — Mohit Shah, Sep 29 '17 at 22:56
Your use of `mapply` may fail without `SIMPLIFY=FALSE` when it returns an `array` vice a `list`; though this only happens if all result in the exact same dimensions, it's still worth guarding against. — r2evans, Sep 29 '17 at 23:07

score 0 · Answer 3 · answered Mar 13 '18 at 01:56

0

You can use mapply to solve the problem

df=mapply(c,vendor_list,problemtype_list) which gives you a data.frame containing values from lists

answered Mar 13 '18 at 01:56

hwq729

3
7

Combine two lists with different structure in R

3 Answers3