fix: drop=FALSE to prevent crash when exactly 1 row is unique to one side

[vertesy]

What was the bug?

combine.matrices.by.rowname.intersect() and merge_numeric_df_by_rn() share the same bug. After computing which row names are unique to each side (diffz[[1]], diffz[[2]] from symdiff()), both do rowSums(matrix1[diffz[[1]], ]) (or the data.frame equivalent) with the default drop = TRUE.

What was the impact?

When exactly one row name is unique to a given side, single-row matrix/data.frame indexing collapses to a plain vector (R's default drop = TRUE behavior), and rowSums() on a non-array throws:

Error in rowSums(...) : 'x' must be an array of at least two dimensions

This isn't a contrived edge case — it's the ordinary shape of "two inputs that share all but one row name each," which is exactly the kind of input these row-merging functions are meant to handle.

The fix

Added drop = FALSE to both subsetting operations in both functions (4 call sites total) — the identical fix for the identical root cause, applied identically.

Testing

# combine.matrices.by.rowname.intersect: 4 rows each, sharing 3, 1 unique per side
m1 <- matrix(1:8, nrow=4, dimnames=list(c("a","b","c","d"), c("x","y")))
m2 <- matrix(9:16, nrow=4, dimnames=list(c("b","c","d","e"), c("x","y")))
combine.matrices.by.rowname.intersect(m1, m2)

# merge_numeric_df_by_rn: same row-name pattern, data.frame inputs
df1 <- data.frame(v1 = c(1,2,3,4), row.names = c("a","b","c","d"))
df2 <- data.frame(v2 = c(9,10,11,12), row.names = c("b","c","d","e"))
merge_numeric_df_by_rn(df1, df2)

Before: both calls errored with 'x' must be an array of at least two dimensions.
After: both complete successfully — the first returns the correctly merged 3x4 matrix, the second the correctly merged 5x2 data frame. Both functions' normal multi-row-difference behavior is unaffected (drop = FALSE is a no-op when more than one row is selected).

• R CMD build + R CMD check: no new NOTEs/WARNINGs/ERRORs vs. the current dev branch (the one remaining ERROR, a missing @export on pU(), is pre-existing and already tracked separately).
• No .Rd/NAMESPACE change needed (roxygen blocks unchanged).
• No testthat/automated tests added, per repo convention.

---

Generated by Claude Code

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: cf7f9a203a

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".